Lit-3913?.
‘ 1325‘!
9a .\ 1.. .

z. :9... 1. . t 5.2: I...
k . . L“. r
.3 . , :1;
25.49 .
. I‘vuﬁrhﬁc’s A . .l 4.1.... .3. V...
.:v. . ‘ . ‘
.x , b

.1»:

3:513...
.24.!) L. 5..
.v: z :

.x :
f y. 7
.2 ..
.3... he.

‘ use.“ .
. y; .9 ..

mm H“.

.l

r: .u u..;\.
’.I‘,> .
1.915» 9
.53-2 .7
.K. . ~n v‘.‘
11.33....11 ..
x ‘ ... .

x
.F,

 

g”..uunz

ﬁ’

 

. . .
, - ‘ v,” ‘
am.“

 

 

”5“ WWiiiiﬁi‘i‘iﬁliﬁiiiiﬂﬁiiﬂ

3 1293 01417 2625

This is to certify that the

dissertation entitled

Relationships Between Treatment Process
Characteristics and Outcome Among
Alcoholic Families in an Early
Child Intervention Program

presented by
Cynthia Lynn Nye

has been accepted towards fulfillment
of the requirements for

Ph .D . degree in Psychology

Majﬂofessor

Date June 13, 1995

 

MSU is an Affirmative Action/Equal Opportunity Institution 0-12771

 

.‘

a v—— ‘...—— ~

 

LIBRARY
Mlchlgan State
Unlverslty

 

 

 

PLACE ll REWRN BOXto remove this checkout from your record.
To AVOID FINES return on or baton duo duo.

DATE DUE DATE DUE DATE DUE

      

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

RELATIONSHIPS BETWEEN TREATMENT PROCESS CHARACTERISTICS
AND OUTCOME AMONG ALCOHOLIC FAMILIES
IN AN EARLY CHILD INTERVENTION PROGRAM
By

Cynthia Lynn Nye

A DISSERTATION

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

DOCTOR OF PHILOSOPHY

Department of Psychology

1995

ABSTRACT
RELATIONSHIPS BETWEEN TREATMENT PROCESS CHARACTERISTICS
AND OUTCOME AMONG ALCOHOLIC FAMILIES
IN AN EARLY CHILD INTERVENTION PROGRAM
By

Cynthia Lynn Nye

Research on parent training interventions has identiﬁed high levels of such family
stressors as marital conﬂict, parent psychopathology, and socioeconomic disadvantage as being
detrimental to treatment effectiveness. Only recently has this literature examined the role of
treatment process characteristics as they relate to parent training outcome, and to pretreatment
family disrupter variables. The present study examined the interplay between parent treatment
investment, parent and therapist expectations, and parent and therapist satisfaction in predicting
change in child behavior and parenting style among families participating in an early intervention
program designed to reduce the development of antisocial behavior problems and later substance
abuse among preschool-aged sons of alcoholic fathers. Parent expectations at pretreatment were
observed to positively inﬂuence early parent investment in the intervention protocol, which in
turn directly predicted child and parenting outcomes. Parent and therapist satisfaction ratings
during treatment were positively associated with parent and therapists’ expectations that the
program would continue to promote change in the target child. Further, treatment investment
and satisfaction were signiﬁcantly correlated, but not predictive of one another with prior levels
controlled. Parent investment across the span of the intervention was identiﬁed as a particularly
salient inﬂuence on outcome, as only families who remained more invested throughout the
intervention showed signiﬁcant improvement in child behavior and parenting at termination.
Baseline parent psychopathology, although correlated with child and parent outcomes, was

generally identiﬁed as a more distal inﬂuence than treatment process characteristics.

ACKNOWLEDGEMENTS

In acknowledging the contributions of those who have facilitated the completion of this
work, I would ﬁrst like to thank Robert Zucker, who has served as an advisor, chairperson,
mentor, and friend. His theoretical expertise and practical guidance have been vital to the
conceptualization of this project, and of past research endeavors. Further, his dedication and
commitment to the progress of his students have been instrumental in promoting my personal and
professional growth over the past several years.

I wish next to recognize the assistance given by Susan Frank, who served as both an
advocate and a resource during the initial stages of this project. In particular, I appreciate her
help in molding my sometimes lofty ideas into more manageable research questions. Thanks are
also extended to Michael Lambert and Thomas Reischl, whose valuable feedback assisted in the
reﬁnement of the current work. Special appreciation goes to Hiram Fitzgerald, who agreed
without hesitation to become involved with this study on very short notice.

On a larger scale, this work would not have been possible without the efforts of Drs.
Zucker, Fitzgerald, and N011 in creating the MSU Longitudinal Study; the exceptional work of
Susan Reﬁor in coordinating the many components of the project; and the many students who
were instrumental in amassing the wealth of information contained in the data base. I am
especially grateful to participating families, who allowed us to share in so many aspects of their
lives.

Finally, I would like to acknowledge the personal support provided by friends and loved
ones. In particular, I am deeply grateful to Kathy for her patience, love, tolerance, and
encouragement throughout all phases of my professional development, and for helping me to

successfully close this chapter of my life.

This work was supported in part by grants to R. A. Zucker, R. B. N011, and H. E.
Fitzgerald from the National Institute on Alcohol Abuse and Alcoholism (AA 07065); the
Michigan Department of Mental Health, Prevention Services Unit; and from the Governor’s
Initiative Against Alcohol and Drug Abuse, Ofﬁce of Substance Abuse Services, Michigan

Department of Public Health.

ii

TABLE OF CONTENTS

LIST OF TABLES
LIST OF FIGURES
CHAPTER I: INTRODUCTION AND REVIEW OF THE LITERATURE
Risk Factors for Later Substance Abuse
The Development of Child Antisocial Behavior
Disrupters of Parenting
Marital conﬂict
Parent psychopathology
Parent Alcoholism, Child Behavior, and Later Substance Abuse
Typologies of Alcoholism
Alcohol-speciﬁc and Non-alcohol-speciﬁc Inﬂuences
The Question of Multiple Risk
The Prevention of Alcoholism
Parent Training Approaches to Child Antisocial Behavior
Parent Training and Substance Abuse Prevention
Parent and Family Disrupters in the Treatment of
Child Antisocial Behavior
Treatment Investment and Treatment Outcome
Predictors of Treatment Investment: Dropout Studies
Predictors of Treatment Investment: Treatment Process Studies
Other Treatment Process Characteristics

Summary

iii

vii

ix

10

13

14

16

24

25

28

29

30

34

37

4O

41

48

Statement of the Problem
Hypotheses
CHAPTER 11: METHOD

Participants

Intervention Procedures

Data Collection Procedures

Measures
Pretreatment Measures: Parent Psychopathology
Pretreatment Characteristics: Family Resources
Pretreatment Characteristics: Child Behavior and Parenting
Treatment Process Measures
Treatment Outcome Measures

CHAPTER III: RESULTS

Data Analytic Procedures
Changes in Child Behavior and Parenting Over Time
Relationship Between Child Behavior and Parenting Over Time

Inﬂuence of Pretreatment Characteristics
Pretreatment Characteristics and Treatment Outcome Measures
Pretreatment Characteristics and Initial Treatment Expectations
Pretreatment Characteristics and Treatment Process Measures

Relationships Among Treatment Process Characteristics
Pretreatment Expectations and Subsequent Investment
Treatment Investment and Parent and Therapist Satisfaction

Parent and Therapist Satisfaction

iv

52

55

57

6O

65

65

66

68

68

71

74

76

78

78

80

80

82

85

86

86

92

Relationships Among Treatment Process and Treatment Outcome Variables
Treatment Satisfaction and Child Behavior Change
Treatment Satisfaction and Changes in Parenting
Parent Treatment Investment and Child Behavior Change
Parent Treatment Investment and Changes in Parenting

Parent, Family, and Treatment Process Characteristics in the Prediction
of Treatment Outcome
Change in Positive and Negative Child Behavior at Termination
Change in Authoritative Parenting at Termination
Tests of Mediation
Power Analyses

CHAPTER IV: DISCUSSION

Pretreatment Inﬂuences on Outcome

Interplay Between Treatment Process Characteristics and Treatment Outcomes
Treatment process characteristics as proximal inﬂuences on outcome
Implications for the treatment process model
Characterizing the ﬂow of treatment

Unique Aspects of the Study

Methodological Limitations

Summary and Future Directions
Implications for treatment

APPENDICES

Appendix A: Pearson Correlations Between Parent Psychopathology
Variables and Psychopathology Clusters

Appendix B: Pearson Correlations Among Family Demographic

V

92

93

93

95

99

102

104

105

105

110

111

115

120

121

122

123

125

129

129

139

and the Family Resources Cluster 140

Appendix C: Family Demographic and Parent Psychopathology
Characteristics of the Treatment Sample 141

Appendix D: Factor Loadings of Observer Impressions Inventory Items in the
Authoritative Parenting Scale Based on Principal Components
Factor Analysis 142

Appendix E: Pearson Correlations Between Selected Revised Belsky Ratings
of Parenting Behavior and Authoritative Parenting Items
on the Observer Impressions Inventory for the Larger
Treatment Sample 143

Appendix F: Pearson Correlations of Parent and Therapist Satisfaction
Ratings With Parent and Therapist Expectations 144

Appendix G: Pearson Correlations Among Mother, Father, and Therapist
Ratings of Child Behavior, and the Composite Child Behavior Scores 146

Appendix H: Tests of Parent Treatment Investment as a Moderator Between
Pretreatment and Post-test Child Behavior and Parenting 148

LIST OF REFERENCES 159

vi

LIST OF TABLES

TABLES

1.

10.

11.

Post hoc Comparisons (Duncan) of Changes in Positive and Negative
Child Behavior and Authoritative Parenting

Partial Correlations Between Pretest (Pretreatment) Parent and Family
Characteristics and Post-test (PT) Positive and Negative Child Behavior,
Controlling for Baseline Positive and Negative Child Behavior Means

Pearson Correlations Between Early and Late Parent Treatment Investment

and Parent and Therapist Satisfaction Ratings at Mid-treatment
and Termination

Hierarchical Regression Analysis of Early (Phase 1) Parent Treatment
Investment as a Predictor of Mid—treatment Parent Satisfaction, With
Pretreatment Parent Treatment Expectations Controlled

Pearson Correlations between Parent and Therapist Satisfaction Ratings
at Post-test Time Points

Partial Correlations (One-tailed) Comparing Parent and Therapist
Satisfaction Ratings with Child and Parent Outcomes at the Same
Time Points, Controlling for Baseline Child Behavior and Parenting Style

Hierarchical Regression Analyses of Parent Treatment Investment as a
Predictor of Positive Child Behavior from Baseline (Pretreatment) to
Mid-treatment, Termination, and Follow-up Levels

Hierarchical Regression Analyses of Parent Treatment Investment as a
Predictor of Negative Child Behavior from Baseline to Mid-treatment,
Termination, and Follow-up Levels

Hierarchical Regression Analysis of Early (Phase 1) Parent Treatment
Investment as a Predictor of Mid-treatment Negative Child Behavior,

With Pretreatment Parent Treatment Expectations and Negative Child
Behavior Controlled

Hierarchical Regression Analyses of Parent Treatment Investment as a
Predictor of Authoritative Parenting from Baseline (Pretreatment) to
Mid-treatment and Termination Levels

Hierarchical Regression Analysis of Parent Psychopathology and

Treatment Process Characteristics as Predictors of Positive Child
Behavior from Baseline (Pretreatment) to Termination Level

vii

PAGE

79

83

88

90

94

96

98

100

101

103

106

12.

13.

14.

15.

16.

17.

18.

19.

20.

Hierarchical Regression Analysis of Parent Psychopathology and
Treatment Process Characteristics as Predictors of Negative Child
Behavior from Baseline (Pretreatment) to Termination Levels

Hierarchical Regression Analysis of Parent Psychopathology and
Treatment Process Characteristics as Predictors of Authoritative
Parenting Style from Baseline (Pretreatment) to Termination

Pearson Correlations Between Parent Satisfaction Items and Parent
Expectations at Post-tests

Pearson Correlations Between Therapist Satisfaction Items and
Therapist Expectations at Post-tests

Pearson Correlations Among Mother, Father, and Therapist Ratings of
Positive Child Behavior, and the Composite Positive Child
Behavior Score

Pearson Correlations Among Mother, Father, and Therapist Ratings of
Negative Child Behavior, and the Composite Negative Child
Behavior Score

Hierarchical Regression Analyses of Parent Treatment Investment as a
Moderator of Pretreatment Positive Child Behavior on Positive Child
Behavior at Termination

Hierarchical Regression Analyses Of Parent Treatment Investment as a
Moderator of Pretreatment Negative Child Behavior on Negative Child
Behavior at Follow-up

Hierarchical Regression Analyses of Parent Treatment Investment as a

Moderator of Pretreatment Authoritative Parenting on Authoritative
Parenting at Termination

viii

107

108

144

145

146

147

150

152

154

LIST OF FIGURES

FIGURE
1. Proposed treatment process model (simpliﬁed) regarding the
interplay between pretreatment and treatment process characteristics

in predicting treatment outcomes

2. Relationship between Pretreatment and Termination (T2) Positive Child
Behavior based on Cumulative Parent Treatment Investment

3. Relationship between Pretreatment and Follow-up (T3) Negative Child
Behavior based on Cumulative Parent Treatment Investment

4. Relationship between Pretreatment and Termination (T2) Authoritative
Parenting based on Cumulative Parent Treatment Investment

ix

PAGE

54

151

155

156

CHAPTER I
INTRODUCTION AND REVIEW OF THE LITERATURE

With an estimated 15.3 million adults in the United States suffering from alcohol abuse
or dependence (US. Dept. of Health & Human Services, 1993), up to 28 million children (West
& Prinz, 1987) are thought to be growing up with at least one parent who is not only alcoholic,
but who may also have problems with depression, antisocial behavior, and other psychological
disturbances associated with impaired parenting (Davies, Zucker, Noll, & Fitzgerald, 1989;
Stoneman, Brody, & Burke, 1989) and poor child outcomes (Farrington & West, 1975; Jacob
& Leonard, 1986; Lahey, Piacentini, McBumett, Stone, Hartdagen, & Hynd, 1988; McCord,
1979; Merikangas, Prusoff, Pauls, & Leckman, 1985; Richman, Stevenson, & Graham, 1982;
Robins, 1966; Robins, West, & Herjanic, 1975; Schneider, Sullivan, Bruckel, Fitzgerald,
Zucker, & Noll, 1989; Stewart, DeBlois, & Cummings, 1980; Stewart & Leone, 1978). Such
statistics suggest that a substantial number of children in these homes are themselves at risk for
psychological impairment. In addition to heightened susceptibility for the development of
alcoholism in later life, children with an alcoholic parent frequently exhibit symptoms of
hyperactivity, conduct disorder, oppositional behavior, and delinquency (e. g. West & Prinz,
1987; Zucker & Noll, 1987) -- all of which have been identiﬁed as precursors to alcoholism and
antisocial behavior in adulthood (Pihl & Peterson, 1991; Zucker, Fitzgerald, & Moses, in press).
Further, research on children of alcoholics indicates that sons of alcoholic fathers are at
particularly high risk for both early conduct problems, and later antisocial behavior and
alcoholism. For example, an estimated 25 to forty percent of these sons are expected to become
alcohol abusers in adult life (Zucker & Noll, 1987). Both the magnitude of this estimate and the
prevalence of precursive behavioral problems among male children of alcoholic fathers argue for
targeting this group Of children for early intervention before problems with alcohol are evident.

Such efforts, aimed at preventing or reducing the early behavioral correlates of later problem

2

drinking, may help to break the intergenerational transmission of alcoholism that frequently
occurs in alcoholic families.

The present research focuses on one such program, the Michigan State University
Multiple Risk Child Outreach Program (Zucker & Noll, 1987), which utilized an outreach
protocol to contact and treat a group of families with preschool age sons, all of whom come from
families with an alcoholic father. Initial analyses (Maguin, Zucker, & Fitzgerald, 1994)
demonstrated that the program was successful in decreasing antisocial behavior and increasing
prosocial behavior among these children both at mid-treatment and at the end of the program.
However, evidence for sustained treatment gains was not observed during the six-month follow-
up in regard to changes in negative child behavior. In addition, this initial outcome study
identiﬁed variation in the extent to which families beneﬁtted from the intervention, relative to
whether one or both parents participated in the program. In order to test for potential differences
in child outcome associated with one- versus two-parent treatment involvement, the protocol
design included two conditions - one in which both parents participated in the program, and one
in which only the mother attended treatment sessions. Some differences in outcome were in fact
found, as families in which both parents participated in the intervention experienced greater
success in increasing their son’s prosocial behaviors than did families in which only the mother
participated.

Overall, then, the program was generally effective in promoting both immediate and
longer-term improvement in the target child’s behavior, although some variation in outcome was
noted. Subsequent research (Nye, Zucker, & Fitzgerald, in press) examined family
characteristics and treatment outcome more closely, and identiﬁed additional differences in the
extent to which families experienced success with the treatment program. Although factors such

as parent psychopathology and family resources were associated with the degree to which child

3

behavior became more positive and less negative by the end of the intervention, maternal
treatment investment was identiﬁed as a critical variable in accounting both for observed variation
in child outcome among treatment families, and for the relationship between pretreatment
characteristics and treatment outcome. Speciﬁcally, signiﬁcant improvement in child behavior
by the end of the intervention was found only among those families who completed the program,
and in which the mother cooperated with homework assignments and was more highly invested
in the success of the regimen. In addition, maternal psychopathology was observed to affect
treatment outcome indirectly, through a strong association with maternal treatment investment
(Nye, 1992). Finally, maternal treatment investment was found to moderate the deleterious
impact of low family resources on treatment success, such that economic deprivation did not
impair the improvement of children whose mothers were more highly invested in the program.

Together, these ﬁndings indicate that treatment process characteristics, such as parent
investment, are critical to an understanding not only of treatment outcome, but also of the impact
of more distal parent and family characteristics on treatment effectiveness. However, only a few
studies on family-based interventions have examined the inﬂuence of parent treatment compliance
or investment on treatment outcome. Instead, this research has tended to focus on pretreatment
parent, child, and family characteristics that are associated with subsequent treatment success.
Although pretreatment characteristics have been linked with outcome, Patterson and Chamberlain
(1988) have provided evidence that parent and family trouble do not directly predict outcome,
but represent more distal disrupters to therapeutic interventions. From this perspective,
pretreatment characteristics inﬂuence outcome by affecting the extent to which parents commit
to and cooperate in parent training programs, which then impairs the acquisition of targeted child-
rearing skills. Thus, it is parent cooperation or compliance that is most directly related to

treatment outcome, through its association with change in parenting practices. In fact, Patterson

4

and his colleagues (e. g. Patterson, Chamberlain, & Reid, 1982) have demonstrated that higher
levels of marital conﬂict and parent psychopathology and fewer family resources interfere with
parent training by increasing client resistance to the treatment and disrupting the acquisition and
application of appropriate child rearing practices.

In light of the identiﬁed primacy of treatment investment to the effectiveness of a variety
of treatment programs and the relative paucity of treatment process research on parent training
interventions, the present study was conducted in order to further examine the role of treatment
process characteristics in accounting for outcome in a parent skills training program. The study
focuses on families participating in the Michigan State University Multiple Risk Child Outreach
Program (Zucker & Noll, 1987). This study represents a more in-depth exploration of treatment
processes and outcomes among the subset of families who completed the intervention protocol.
The goal of this work is to further understand how different aspects of treatment are involved in
leading to better or worse outcomes. It was anticipated that such process characteristics as parent
investment, parent and therapist expectations, and parent and therapist satisfaction, would
represent direct or proximal inﬂuences on outcome, whereas pretreatment characteristics such as
parent psychopathology and family socioeconomic resources would be more distally associated
with treatment success.

The following sections describe the empirical support for the preventative approach used
in the MSU. program, including an examination of the linkages between child antisocial
behavior and later substance abuse, a review of characteristics which place male sons of alcoholic
fathers at elevated risk for both early conduct problems and later difﬁculty with alcohol and other
drugs, and a description of child- and family-focused substance abuse prevention approaches.
Research on the inﬂuence of pretreatment characteristics on treatment outcome is then presented,

as these were considered to be potentially important but more distal determinants of treatment

5

success in the current work. Finally, literature documenting the more proximal inﬂuence of
treatment process characteristics on outcome is reviewed, with a particular emphasis on the
relationships between treatment investment, expectations, and satisfaction in predicting outcome
in family-based, parent skills training interventions. The study itself examines how these
treatment process characteristics are associated with one another and with treatment outcome
among the alcoholic families completing the M.S.U. Program, with pretreatment child, parent,
and family characteristics controlled.
Risk Factors for Later Substance Abuse

Substantial research has been conducted in efforts to identify precursors to alcohol and
other drug abuse in adolescence and early adulthood. Among the various risk factors, several
characteristics of child and family functioning have been associated with later substance abuse.
These factors include: 1) family alcohol use and attitudes, 2) poor and inconsistent disciplinary
practices, 3) marital conﬂict, and 4) early and chronic antisocial child behavior (Hawkins,
Catalano, & Miller, 1992).

Regarding child behavior, there is substantial evidence for a link between early antisocial
behavior and both later antisocial behavior and substance abuse. Aggressive conduct problems
displayed in childhood are often early signs of more severe antisocial behavior later on, and many
researchers have pointed out the continuity of antisocial behaviors from childhood into
adolescence and adulthood (Eron, Huesmann, Dubow, Romanoff & Yarmel, 1987; Gersten,
Langer, Eisenberg, Simcha-Fagan & McCarthy, 1976; Loeber & Dishion, 1983; Windle, 1989).
Furthermore, the earlier and more extreme such problems are, the more severe and chronic tends
to be the behavior pattern in adolescence and adulthood (Loeber, 1986; Tolan & Lorion, 1988).
From this perspective, an early onset of antisocial behavior is particularly predictive of

continuity, and is also indicative of the severity of later antisocial acts. In addition, there is

6

substantial evidence that early antisocial behavior is predictive of adolescent substance abuse
(Johnston, O’Malley, & Evelard, 1978; Kandel, Kessler, & Margulies, 1978; Robins, 1978;
Wechsler & Thum, 1973).

Loeber (1986) has proposed that the developmental progression and outcome of antisocial
behavior is largely determined by when and to what degree early conduct problems appear.
According to this formulation, child antisocial behavior progresses over time from acts of minor
severity to ones that are increasingly serious, with the worst outcomes including violent offenses
and polydrug abuse. Further, this developmental pathway includes earlier initiation into alcohol
use, which has been associated with subsequent misuse of alcohol (e. g. Rachal, Guess, Hubbard,
Maisto, Cavanaugh, Waddell, & Benrud, 1982), and with greater involvement in other drug use
(e.g. Fleming, Kellam, & Brown, 1982; Kandel, 1982).

As corroborated by other research (e.g. Collins, 1981; Halikas & Rimmer, 1974;
Harwood & Leonard, 1989; Heather, 1982; Hesselbrock, Hesselbrock, & Stabenau, 1985;
Kandel, 1982; Lewis, Moy, Jackson, Aaronson, Restifo, Serra, & Sirnos, 1985; Loeber, 1982;
Mills & Noyes, 1984; Robins, 1966), early onset of child antisocial behavior is highly predictive
of later delinquency, sociopathy, and substance abuse. Because of the strength of these
relationships, early antisocial behavior has been identiﬁed as a potential marker for later
substance abuse and, indeed, has been so identiﬁed in the present work, as well. In the
following sections, these areas of risk are reviewed in more detail.

The Development of Child Antisocial Behavior

An extensive body of research indicates that parental disciplinary practices contribute
substantially to the development of aggressive child behavior and adolescent delinquency (e. g.
Cerkovich & Girodano, 1987; Lahey, et al., 1988; VanVoorhis, Cullen, Mathers, & Garner,

1988). In particular, poor supervision and lack of parental involvement in children’s activities

7

have been identiﬁed as the two strongest parenting correlates of severe child conduct problems
(Loeber & Stouthamer-Loeber, 1986).

Other authors have focused on inconsistent discipline as a precursor to antisocial child
behavior (e. g. Patterson & Stouthamer-Loeber, 1984; Wells & Rankin, 1988). This latter model
proposes a coercive process whereby parents respond irritably and/or ineffectively to their young
child’s coercive and aggressive behavior. Because of their ineffective discipline and monitoring
of antisocial behavior, the parents are not able to maintain an appropriate level of compliance
(Patterson & Bank, 1987). Over time, both the aggressiveness of the child and the ineffectiveness
of the parents increase and serve to maintain one another. As a result, the child learns and
develops this coercive pattern of behavior as a way of dealing with others, and later generalizes
his antisocial interactional style to classroom and playground settings. In each new setting, the
child’s noncompliance and aggressiveness are responded to with aggression and rejection from
others, thus recreating the negative interactions taking place in the home. As these cycles
continue, the child’s failure to acquire positive social skills, coupled with his coerciveness,
"creates high risk for afﬁliation with delinquent peers, [thereby] further enhancing the
development of chronic antisocial behavior" (Reid & Patterson, 1988, p. 115).

Other researchers (e. g. Dadds, 1987; Wahler & Dumas, 1987) also highlight the role that
coercive parent-child interactions play in the development of antisocial behavior problems in
children. According to Dadds (1987), for example:

Although a diversity of factors may be associated with the development and

maintenance of conduct/oppositional disorders in children, of primary importance

are the moment-to—moment interactions that the child has with his/her primary

caregivers. These are often marked by coercive, aggressive behaviors that may
be functional for parents and children within the family system (p. 341).

Disrupters of Parenting

Given the apparent primacy of parenting factors to the development of child antisocial
behavior, it is important to identify parent and family characteristics that impede parents’ ability
to adequately monitor and engage with their children. In this regard, both parent
psychopathology (e.g. alcoholism, depression, and antisocial behavior) and marital conﬂict or
discord have been implicated. Patterson and his colleagues have proposed a mediational model,
whereby contextual variables such as parent psychopathology, stress, and economic disadvantage
disrupt family processes (Patterson, Reid, & Dishion, 1992). Thus, these parent and family
characteristics contribute to antisocial behavior in children to the extent that they disrupt day-to-
day parenting practices and contribute to ineffective discipline and poor supervision. In fact,
Reid and Patterson suggest that any such condition that disrupts or impedes the daily socialization
of the child indirectly contributes to the subsequent development of child antisocial behavior.
From this perspective:

...although many factors such as parent criminality, social and economic

disadvantage, child temperament, and marital discord systematically affect the

development of antisocial behavior, their inﬂuence is mediated by the extent to

which they disrupt day-to-day parenting practices. Particularly... irritable,

ineffective discipline and poor parental monitoring are the most proximal

determinants of the early development and maintenance of antisocial behavior

(Reid & Patterson, 1988, p. 126).

Similarly, Dadds (1987) states that the likelihood that parents will engage in coercive interactions
with their children is related, in part, to marital satisfaction and social support characteristics.
Thus, parent psychopathology, marital conﬂict, and economic deprivation may affect parents’
child-rearing abilities which, in turn, contribute to antisocial behavior in the child. Others (e. g.
Belsky, 1984; Belsky, Robins, & Gamble, 1984) have also described parental functioning as

being multiply determined, with individual, interpersonal, and contextual factors all playing an

important role.

9
Marital conﬂict. Belsky (1981) has proposed that the marital relationship is the principal

support system for parents (especially mothers), and several studies have linked marital conﬂict
to the development of child behavior problems (Emery & O’Leary, 1982; Jaffe, Wolfe, Wilson,
& Zak, 1986; Johnson & O’Leary, 1976; Jouriles, Murphy, & O’Leary, 1989; McCord, 1979;
Porter & O’Leary, 1980; Richman, Stevenson, & Graham, 1982; Wolfe, Jaffe, Wilson, & Zak,
1985). Research has also identiﬁed an association between marital discord and particular
disciplinary strategies. For example, a high incidence of both interspousal aggression and parent-
to.child aggression has been found among alcoholic families (e. g. Dielman, Barton, & Cattell,
1977; Ellwood, 1980; Famularo, Stone, Barnum, & Wharton, 1986; Leonard, Bromet,
Parkinson, Day, & Ryan, 1985; Reider, Zucker, Noll, Maguin, & Fitzgerald, 1988).

Other research lends additional support to the link found between marital quality and
parenting behavior (e. g. Brody, Pillegrini, & Sigel, 1986; Frank, Hole, Jacobson, Justkowski,
& Huyck, 1986). Marital satisfaction has been found to be positively related to more positive
and less punitive parenting (Johnson & Lobitz, 1974; Kemper & Reichler, 1976). Gottrnan and
Katz (1989) found that maritally distressed couples displayed a cold, unresponsive, and angry
parenting style that was low in limit setting and structuring. Further, this parenting style was
associated with angry and noncompliant behavior in the young children. Similarly, Jouriles and
his colleagues (Jouriles, Pﬁffner, & O’Leary, 1988) observed an association between marital
dissatisfaction and both parenting style and child deviance among a group of mothers and
toddlers. Further, conﬂictual marital exchanges have been found to inﬂuence subsequent parent-
son interactions, with sons observed to be less compliant with their father, and with fathers found
to issue more vague or confusing commands toward their son at the beginning of the observation
session (J ouriles & Farris, 1992). Particularly noteworthy in this last study is the fact that the

sons had not witnessed the conﬂict between the parents prior to interacting with their fathers, but

10

were nonetheless more oppositional than sons whose parents had not previously had a conﬂictual
interaction.

Fauber and his colleagues (Fauber, Forehand, Thomas, & Wierson, 1990) examined the
extent to which the relationship between marital conﬂict and the psychological adjustment of
young adolescents is mediated by parenting behavior. Parenting characteristics were observed
to be directly associated with child internalizing symptoms, with marital conﬂict only associated
via its impact on parenting. Parent marital conﬂict was observed to have both direct and indirect
inﬂuences on externalizing symptoms among the sample of adolescents whose parents were not
divorced, but only an indirect effect via parenting for adolescents whose parents were divorced.
The authors propose that marital conﬂict may contribute to child acting-out behaviors both by
impairing parenting practices and by exposing children to confrontational or aggressive
interactions.

Parent psychopathology. Parent alcoholism, antisociality, and depression have all been
associated with child antisocial behavior (e. g. Farrington & West, 1975; Jacob & Leonard, 1986;
Lahey, et al., 1988; McCord, 1979; Merikangas, et al., 1985; Reider, Zucker, Maguin, Noll,
& Fitzgerald, 1989; Richman, et al., 1982; Robins, 1966; Robins, et al, 1975; Schneider, et al.,
1989; Stewart, et al., 1980; Stewart & Leone, 1978). However, few studies have directly
examined the direct inﬂuence of these disorders on speciﬁc child rearing practices. Kuczynski
(1984) has suggested that parents who are experiencing distress or depression choose more
automatic and less cognitively taxing discipline strategies, rather than more purposive and
effortful techniques, such as rational guidance. Others (e. g. Brody & Forehand, 1986) propose
that parental depression and other dysfunctions in parental adjustment can decrease the parent’s
tolerance for noncompliance, thereby making even innocuous behaviors seem bothersome. This,

in turn, may lead to a greater use of punishment and authoritarian control at a time when the

11

parent is typically trying to limit his/her interactions with the child. The result is the
development of the coercive style of parent-child interactions thought to underlie child antisocial
behavior (e.g. Patterson, 1982). *

These theoretical formulations have been supported by observational research on
depressed women (Hops, Biglan, Sherman, Arthur, Friedman, & Osteen, 1987), as interactions
between these women and their families have been found to be highly aversive in nature. In
contrast, Stoneman and her colleagues (1989) found both marital and individual distress in men
to be predictive of inconsistent fathering and lack of parental agreement about discipline, while
no such effects were found for women.

In one of the few studies explicitly focused on a sample Of alcoholic parents, Davies and
his colleagues (Davies, Zucker, Noll, & Fitzgerald, 1989) examined the relationship between
several types of parental psychopathology and child-rearing practices among young alcoholic
families in the Michigan State University Longitudinal Study. More speciﬁcally, the researchers
studied the association between self-reported parenting practices and parent symptomatology for
alcoholism, depression, and antisocial behavior. For the alcoholic fathers, problems in the
reported affective relationship with the child were strongly associated with the father’s self-
reported depression, as well as with his current level of alcohol consumption and his lifetime
alcohol problems. Further, the antisocial behavior of the fathers was related to inconsistent
discipline and an increased demand for the child to be independent. Other studies with the same
group of alcoholic fathers (Fitzgerald, et al., 1989; Reider, et al., 1989; Zucker, Weil, Baxter,
& Noll, 1984) conﬁrm the inﬂuence of paternal antisocial behavior on father-child interactions
and on child behavior problems.

Among the mothers in the Davies (et al., 1989) study, both antisocial behavior and

depression were associated with a disrupted parent-child affective relationship. In addition,

l2

paternal depression had a deleterious effect on maternal child rearing. This corroborates other
research with these families (Schneider, et al., 1989) which found that paternal depression may
indirectly affect child outcome by increasing maternal depression and impairing mothers’ child-
rearing practices. It also underscores the inﬂuence that parental psychopathology can have on
speciﬁc child rearing behaviors among alcoholic families.

Observations of parent-child interactions among families in the M.S.U. study revealed
some differences in psychopathology and parenting behavior among parents in alcoholic versus
nonalcoholic families. Parents in the alcoholic families had more extensive histories of antisocial
behavior, and demonstrated greater difﬁculty in engaging their children. In addition, these
parents displayed less warmth and affection toward their children during a clean-up session than
did nonalcoholic parents (Whipple, Fitzgerald, & Zucker, 1995).

Chassin and her colleagues (Chassin, Pillow, Curran, Brooke, Molina, & Barrera, 1993)
found that maternal alcoholism was related to impaired monitoring of adolescents’ activities by
both mothers and fathers (which, in turn, was related to adolescents’ substance use and afﬁliation
with drug-using peers). Fathers’ monitoring was further affected by level of parent antisociality.

Thus far, this review of the literature has focused on the interconnections between child
antisocial behavior, parenting practices, marital conﬂict, and parent psychopathology - all of
which have been associated with the development of substance abuse (Hawkins, et al., 1992).
However, given the evidence that children of alcoholics are at elevated risk for alcoholism as
adults (e. g. Sher, Walitzer, Wood, & Brent, 1991; West & Prinz, 1987), the relationship between
parent alcoholism and child substance abuse needs to be more explicitly described. The following
section focuses on what is currently known about parent alcoholism as a risk factor in the

development of child antisocial behavior and later substance abuse.

13
Parent Alcoholism Child Behavior and Later Substance Abuse

 

The impact of alcoholic parents on the functioning of children has been well documented
in the psychological literature (e. g. Adler & Raphael, 1983; Lord, 1983; Roosa, Sandler, Beals,
& Short, 1988; Rydelius, 1981; Seixas, 1977; West & Prinz, 1987). Such studies have identiﬁed
a relationship between antisocial child behavior and parental alcoholism, and male offspring in
particular have been found to be at risk for conduct disorder and antisocial behavior in childhood,
substance use and delinquency in adolescence, and alcoholism and antisocial personality as adults
(e.g. Zucker & Noll, 1987).

Although the research indicates that children of alcoholics are at increased risk for
developing problems with alcohol themselves (e. g. Bennett & Wolin, 1985; Cotton, 1979;

Goodwin, 1985), the mechanisms by which risk evolves into actual trouble require further

 

explication. Thus far, the impact of parent alcoholism on children has been explained in a
number of ways, including genetics, modeling, and other family inﬂuences:

Beyond the genetic transmission of a propensity to alcoholism in males, family

modeling of drug using behavior and parental attitudes toward children’s drug

use are family inﬂuences related speciﬁcally to the risk of alcohol and other drug

abuse. Poor parenting practices, high levels of conﬂict in the family, and a low

degree of bonding between children and parents appear to increase the risk for

adolescent problem behaviors generally, including the abuse of alcohol and other

drugs (Hawkins, et. al., 1992, p. 82).

Although male children of alcoholic fathers seem to be at disproportionate risk for
alcoholism, there is also evidence that these "high risk" offspring do not uniformly develop
problems with alcohol (Alterman, Bridges, & Tarter, 1986; KnOp, Teasdale, Schulsinger, &
Goodwin, 1985; Schuckit & Sweeney, 1987). Thus, it is necessary to examine variation among
alcoholic families that may place children at greater or less risk for substance abuse problems.

As noted in the study by Chassin and her colleagues (Chassin, et al., 1993), for example, risk

for substance abuse among early adolescent children of alcoholics was directly related to parent

14

levels of alcohol consumption, and more indirectly inﬂuenced by the impact of parent alcoholism
and antisociality on parents’ monitoring of their children’s behavior (which, in turn, inﬂuenced
afﬁliation with drug-using peers). Further, the authors noted that more extensive parent
alcoholism was associated with greater environmental stress, the latter of which also predicted
the adolescent substance use and afﬁliation measures.

Typologies of Alcoholism .

There is growing evidence that alcoholism is not a unitary entity, but that its development
and course may follow different trajectories. Cloninger has distinguished between Type II (or
"male limited") alcoholism and Type I (or "milieu limited") alcoholism. Under this formulation,
Type II alcoholism represents an early onset form of alcoholism that is thought to be largely
driven by genetic mechanisms. In contrast, Type I alcoholism develops later, and is associated
more with environmental than genetic factors (e. g. Cloninger, 1987). In terms of diagnostic
differences, Type I alcoholics are thought to suffer more psychological distress about their alcohol
use, whereas Type II alcoholics tend to experience more negative social consequences (e. g. loss
of job, family arguments, etc.) because of their drinking. Type II alcoholics are further
distinguished as having an earlier onset of alcohol abuse, with heavy drinking beginning before
the age of 25.

Although recent research lends some support to these subtypes, it also indicates that even
Type II alcoholism may be associated with risk factors other than family alcoholic history. In
examining different developmental pathways associated with the two subtypes, Zucker and his
colleagues (Zucker, Ellis, & Fitzgerald, 1993a, 1993b) found that an alcoholic rearing
environment (e.g. being raised by an alcoholic parent) was associated with alcoholic severity for
Type I alcoholics, whereas extended family history of alcoholism was associated with later

alcoholic trouble for Type H alcoholics. In addition, Type II alcoholism was associated with both

15

childhood antisocial behavior and adult antisociality. Thus, early conduct-disordered behavior
was predictive of later antisocial behavior and alcohol abuse for Type II alcoholics, beyond the
inﬂuence of alcoholic family background.

Despite ﬁnding some evidence in support for Cloninger’s subtypes, Zucker and his
colleagues (Zucker, et al., 1993b) have proposed an alternative schema based on the association
between antisociality and alcohol problems in adulthood, and on the developmental continuity
found between antisocial behavior in childhood and adulthood. The authors classiﬁed subjects
as either antisocial alcoholics (AALs) or non-antisocial alcoholics (NAALS), based on a median
split of lifetime antisocial behavior. According to this approach, the higher scores of the
antisocial alcoholics "indicate a trajectory which begins early in life with aggressive/antisocial
behavior, and which thereafter crystallizes in alcoholism and antisociality in adulthood" (p. 13).
In examining differences between the two typologies, the authors found that antisocial alcoholics
had an earlier onset of alcohol problems, more extensive alcohol-related difﬁculty, lower
socioeconomic status, greater lifetime and current depression, and a more extensive family history
of alcoholism. In models predictng severity of alcoholism, both child and adult antisocialityI and
family history of alcoholism were signiﬁcant for antisocial alcoholics, while variations in alcohol
difﬁculty for non-antisocial alcoholics were predicted only by lifetime depressive
symptomatology.

Research on adolescent alcohol use onset supports the identiﬁcation of different
developmental pathways related to drinking. Weber and his colleagues (Weber, Graham, Hansen,
Flay, & Johnson, 1989) examined drinking patterns over time among "Type II" adolescents -
who were identiﬁed as being "problem behavior prone", based in part on their apathy toward the
reactions of signiﬁcant others’ toward their alcohol use as well as their greater association with

alcohol-using peers -- and among "Type I" adolescents who the authors describe as being more

16

"normally socialized ". Type II adolescents showed signiﬁcantly higher rates of alcohol use and
drunkenness at follow-up than did the Type I sample, thereby suggesting an accelerated
progression of substance abuse among the problem behavior prone group. Although this study
did not directly measure differences in externalizing or antisocial behavior problems between the
two groups, one might expect a greater degree of antisocialin among the Type II adolescents,
given their greater afﬁliation with delinquent peers and their relative indifference to social cues
about their drinking.

In light of these indications that alcoholism may develop and progress in at least two, if
not more (Zucker, in press) separate pathways, type of parent alcoholism may be differentially
associated with child functioning. Support for this hypothesis may be found in the Zucker
(Zucker, et a1. , in press) study, which identiﬁed a higher level of externalizing behaviors among
preschool age sons of antisocial alcoholics. This ﬁnding presents preliminary evidence that these
young male sons of antisocial alcoholic fathers may themselves be at risk for the development of

antisocial alcoholism.

Alcohol-speciﬁcﬂd Non-alcohol-speciﬁc Inﬂuences

 

In addition to examining differences in alcoholism subtypes as they related to child risk
for substance abuse, research has also attempted to differentiate the inﬂuence of factors
speciﬁcally related to alcohol use and those not directly pertaining to alcohol consumption (e. g.
Zucker, 1993). Zucker and Fitzgerald (1991) emphasize the importance of pursuing both of these
latter areas of inquiry into risk for alcoholism:

One line focuses on ways in which children are exposed to and learn about the

idea that alcohol is a substance that can be used to create a change in how one

feels and how one acts. The other line of investigation examines factors that are

not speciﬁc to alcohol use, but that precede it and are parts of the causal chain

of problem alcohol involvement (p. 19).

Alcohol-speciﬁc risk factors. There is evidence that even very young children can

l7

recognize and identify alcoholic beverages, are aware that alcoholic beverages are special
substances, and recognize that there are cultural norms regarding their consumption (Noll,
Zucker, & Greenberg, 1990; Noll, Zucker, Weil, & Greenberg, 1984). N011 and his colleagues
reported that preschool-age children in their study identiﬁed alcoholic beverages as being adults’
beverage of choice, and tended to select alcohol for men to drink more often than for women
(Noll, et al., 1984). In addition, children who were more successful at recognizing alcoholic
beverages (by smell) tended to come from homes in which there was heavier parental drinking
(Noll, et al., 1990). Thus, children from families in which there is more extensive and
problematic alcohol use seem to be differentially socialized about drinking practices and may
develop expectancies about alcohol at an earlier age (Miller, Smith, & Goldman, 1990; Zucker
& Fitzgerald, 1991; Zucker, Kincaid, Fitzgerald, & Bingham, in press). This ﬁnding is
particularly important in light of evidence indicating that children’s expectations about alcohol
use are related not only to parental patterns of alcohol use, but also to the children’s own
subsequent patterns of use in adolescence (Brown, Creamer, & Stetson, 1983; Christiansen &
Goldman, 1983). Further, research has suggested that initiation into the use of a substance is
preceded by attitudes in favor of its use (Kandel, Kessler, & Margulies, 1978; Krosnick & Judd,
1982; Smith & Fogg, 1978).

There is also evidence that drinking-speciﬁc variables, such as history, frequency,
amount, consequences, and current consumption, have an impact on family relationships and
parent-child interactions. For example, an individual’s history of lifetime alcohol problems is
related to several aspects of functioning in alcoholic families. As described above, Chassin and
her colleagues (Chassin, et al., 1993) found an association between maternal alcoholism and
father and mother monitoring of their adolescent’s activities. In addition, level of father and

mother alcoholism was related to greater environmental stress. Reider and her colleagues

18

(Reider, et al., 1988) examined interspousal aggression among 75 working-class couples from
the Michigan State University Longitudinal Study. Husbands reporting higher levels of violence
towards their wives in the past year were younger, had a more extensive history of drinking
problems, and engaged in more antisocial activity. In those families, wives also were younger
in age and were currently drinking less alcohol than were the wives of husbands reporting lower
levels of violence. Greater lifetime alcohol problems and higher levels of prior antisocial activity
in both men and women were also found to be associated with the number of marital separations.

Regarding the family environment in these homes (as measured by the PBS (Moos &
Moos, 1981)) greater conﬂict was experienced by husbands who had a more extensive history
of alcohol-related problems, and both greater conﬂict and less family cohesion were experienced
by husbands and wives who reported higher levels of violence towards their spouses. In
accounting for the alcoholic husband’s violence towards his wife, the husband’s age and his long-
term drinking problems were the best predictors of the overall level of violence, and the prior
antisocial behavior of the husband was predictive of the more severe forms of violence reported.

In another study, Reider and her colleagues (1989) found that child-directed parental
aggression was positively related to the extent Of antisocial behavior, depression, and long term
alcohol involvement in the male alcoholics and their partners. Further, parental aggressiveness
and lifetime alcohol problems were found to be the most predictive of aggression directed against
the young male children targeted in the study.

These studies suggest, then, that an extensive history of alcohol-related difﬁculties is
associated with physical marital conﬂict, marital separation, and reciprocal aggression between
parents and children -- all of which may place children at risk for substance abuse (Hawkins, et
al., 1992). Similarly, Leonard and his colleagues (1985) found a relationship between reported

history of alcohol-related problems and physical marital conﬂict. Davies and his associates

19

(1989) discovered that child aggression against the parents, and disengaged and inconsistent
fathering were related to the lifetime alcohol problems of both the alcoholic fathers and their
partners. Further, reported problems in the fathers’ affective relationship with the child were
related to self—reported depression, and to both lifetime alcohol problems and current alcohol
consumption.

Several other studies also suggest that the alcoholic’s current level Of consumption is
intimately tied to other aspects of ongoing family relationships. The series of Moos studies
(Moos, Finney, & Chan, 1981; Moos, Finney, & Gamble, 1982; Moos & Moos, 1984),
conducted with a somewhat older sample than the Michigan State project, compared the
functioning of relapsed, recovered, and nonalcoholic individuals and their families. Children of
recovered alcoholics (no longer drinking at the 18 month post-treatment follow-up) were reported
by the parents to be functioning as well as the children in the community control group (Moos
& Billings, 1982). In contrast, children of relapsed alcoholics had more emotional and physical
problems than the children in the other two groups. Further, children in the relapsed group were
living in an environment that was described (on the PBS) as being lower on family cohesion,
expressiveness, parental congruence, and family activities. This suggests, then, that child
functioning is impaired by family difﬁculties related to the alcoholic’s consumption of alcohol.

In other Moos studies (Moos, Finney, & Chan, 1981; Moos, Finney, & Gamble, 1982;
Moos & Moos, 1984) relapsed alcoholics and their families also appeared to be functioning more
poorly than the recovered alcoholic and the nonalcoholic families on a variety of dimensions.
For example, relapsed alcoholics reported more depression, anxiety, and physical symptoms.
Consistent with other research (Filstead, McElfresh, & Anderson, 1981), these individuals also
described their family environment (on the PBS) as being less cohesive, expressive, and

organized, and more conﬂictual than did members of the recovered and nonalcoholic groups

20
(Moos, et al., 1981).

The spouses of these relapsed individuals (Moos, et al., 1982) provided similar reports
of distress. They reported more negative life events and perceived less cohesion in their families
than did spouses of recovered and nonalcoholic individuals. Further, spouses of heavily drinking
relapsed alcoholics were more depressed, more anxious, engaged in fewer informal social
activities, and reported more negative life events than the spouses of relapsed alcoholics who
were. trying to control or reduce their consumption. Spouses of heavy drinkers also perceived
their family as being more conﬂictual, less cohesive and organized, and with less of a recreational
orientation.

These ﬁndings indicate that both the drinking status and the consumption pattern of the
alcoholic partners interconnect with the functioning of the marital partner and with parent
perceptions of the home environment. It is not unreasonable, therefore, to conclude that such
differences at the parental level are likely to inﬂuence these parents’ interactions with their
children. A later study by Moos and Moos (1984), using different groups of alcoholics and
matched community controls, provides further conﬁrmation of this. Not only were families of
relapsed alcoholics described as less cohesive and expressive than were the other families, but
families of heavily-drinking relapsed alcoholics again seemed to be the most disrupted, with more
family arguments, lower cohesion, more conﬂict, and less organization. Regression analyses
determined that:

the families of the [recovered and relapsed] alcoholics were strongly affected by

the level of adaptation of the alcoholic partners. Families in which the alcoholic

members reported more alcohol consumption and drinking problems and

complained of more anxiety, depression, and physical symptoms had more family
arguments, less cohesion and expressiveness, and showed less agreement about

their family environment and about joint task performance. . .Cohesion was [also]

lower in families in which the spouses complained of more anxiety, and

expressiveness was lower in families in which they complained of more
depression...(p. 115).

21
Other research (e.g. Dunn, et al., 1987; Jacob, 1987; Jacob, Ritchey, Cvitkovic, &

Blane, 1981; Steinglass, 1980a, 1980b, 1981, 1987; Steinglass, Tislenko, & Reiss, 1985) has also
found a relationship between current alcohol consumption patterns and disrupted family
interactions and relationships. With regard to parent—child interactions, Seilhamer and Jacob
(1990) generate an important causal hypothesis that attempts to link these ﬁndings:

(a) there is an ongoing association between parental drinking and the parent-child
relationship, (b) this association involves a causal relationship, in that
drinking/ intoxication effects disturbances in the parent-child relationship, and (c)
while parental alcoholism is assumed to cause negative outcomes for children in
the long run, the quality of the parent-child relationship during day-to-day cycles
of sobriety and intoxication may vary with drinking pattern and consumption
level (p. 30).

 

Non-alcohol-speciﬁc risk factors. Research on alcoholics has identiﬁed a high rate of
comorbidity between alcoholism and other forms of psychopathology. For example, the national
Epidemiologic Catchment Area (ECA) Study estimates that 37 percent of the general population
with an alcohol abuse/dependence diagnosis also qualiﬁes for an additional diagnosis not related
to substance abuse (Regier, Farmer, Rae, Locke, Keith, Judd, & Goodwin, 1990). Further, this
rate is even higher (55%) among alcoholics in treatment, suggesting that comorbidity is
particularly prevalent among individuals with severe alcohol involvement. Among the most
frequent comorbid diagnoses for this group are anxiety disorder (19%), antisocial personality
(14%), and affective disorders (13%) (Regier, et al, 1990).

As already shown, parent characteristics not speciﬁcally related to alcohol consumption
can also have an impact on family interactions. In particular, this research has identiﬁed other
types of parent trouble, including antisociality and depression, that affect the home environment.
For example, Reider and her colleagues (1989, 1988) found an association between parent
antisocial behavior and depression and child-directed aggression, and between the prior antisocial

behavior of the parents and physical marital conﬂict among the young alcoholic families of the

22
Michigan State University Longitudinal Study. Chassin and her colleagues (Chassin, et al. , 1993)

found an association between parent antisociality and fathers’ monitoring of their adolescent
child’s activities.

Others (e.g. Merikangas, et al., 1985; Reider, et al., 1989; Schneider, et al., 1989) have
observed that depression in alcoholic parents and/or their spouses is related to child antisocial
behavior. Reider, et al. (1989) found that young male children’s aggression towards their parents
was related to depression in mothers and fathers, as well as to the lifetime alcohol problems of
the parents. In a study focusing on secondary alcoholism in parents with major depression
(Merikangas, et al., 1985), adult offspring of depressed parents had a much higher rate of
antisocial personality than did the adult controls. Similarly, younger children of the probands had
a greater incidence of conduct disorder than did the controls. When parental alcoholism as a
secondary diagnosis was considered, the risk of antisocial behavior and of conduct disorder was
markedly increased. Additionally, offspring of two alcoholic parents had a greater rate of
antisocial personality or conduct disorder than did offspring with only one alcoholic parent.

Because of the frequent co-occurrence of other forms of psychopathology with
alcoholism, it is also important to examine the extent to which risk for alcoholism is associated
with the cumulation of parent psychopathology, or whether each disorder seems to have unique
effects on child functioning. As described above, Davies and his colleagues (Davies, et al.,
1989) examined the effects of parent psychopathology on child-rearing behavior among alcoholic
families in the Michigan State Longitudinal Study. Both paternal depression and alcohol
problems were associated with difﬁculties in the fathers’ affective relationship with the male
target child, while the quality of the mother-child relationship was negatively associated with
maternal depression and antisociality. In addition, paternal antisocial behavior was related to

inconsistent discipline and demands that the child be more independent among the fathers.

23

Similarly, Chassin (Chassin, et al., 1993) found an association between parent antisociality and
impaired monitoring by alcoholic fathers of young adolescent children.

Other research by Chassin and her colleagues (Chassin, Rogosch, & Barrera, 1991)
indicates that symptomatology among children of alcoholics may be related to different aspects
of parent functioning. In this study of a community sample of adolescents and their parents, the
authors identiﬁed an alcoholic sample in which 74.6% of the alcoholic fathers and 58.5 % of the
alcoholic mothers reported an early onset of drinking problems (on or before age 25).
Comorbidity rates were fairly comparable to the ECA data for alcoholic men and women (Helzer
& Pryzbeck, 1988), with a 5.5 % prevalence for major depression and a 16.4% prevalence of
antisocial personality among the men, and rates of 12.1% and 15.5 % of depression and antisocial
personality, respectively, among the women. Overall, the authors noted that the group of
adolescent children of alcoholics was characterized by lower parent education, greater parent drug
use, more environmental stress and family disruption, and higher rates of parent psychopathology
than the control group (Chassin, et al., 1991).

In predicting adolescent symptomatology from parent psychopathology measures, the
Chassin group found that adolescents’ externalizing behavior was related to parent antisociality,
while fathers’ alcoholism was a signiﬁcant predictor of both alcohol use and substance-related
consequences or dependence symptoms among the adolescents. Parent antisocial personality did
not make a unique contribution to the prediction of alcohol use, but was associated with
adolescent drug use. In examining differences among children of current (past 3 years) versus
recovered alcoholics, the authors further noted that externalizing symptomatology was higher
among children of current alcoholic fathers than among those whose fathers were no longer
drinking. Thus, parent alcoholism may differentially affect child antisocial behavior when current

consumption is taken into account. This has been supported by other research indicating higher

24

levels of functioning among recovered versus relapsed alcoholics, and among spouses and
children of the recovered group (Moos, Finney, & Chan, 1981; Moos, Finney, & Gamble, 1982;
Moos & Moos, 1984).

Together, these ﬁndings indicate that different aspects of parent functioning may have
independent effects on both child and family functioning, thereby making unique, as well as
additive, contributions to risk for substance abuse and its sequelae. In particular, this research
identiﬁes parent antisociality as an important predictor of child acting out behavior (Chassin, et
al., 1991; Fitzgerald, et al., 1989; Reider, et al., 1989; Zucker, et al., 1984; Zucker, et al., in
press), parental monitoring and disciplinary practices (Chassin, et al., 1993; Davies, et al.,
1989), and level of marital conﬂict and violence (Reider, et al., 1988, 1989). Parent alcoholism
appears to have a direct inﬂuence on child alcohol use (Chassin, et al., 1991, 1993), in addition
to an association with marital aggression (Reider, et al., 1988, 1989), and disrupted parenting
(Chassin, et al., 1993; Davies, et al., 1989; Whipple, et al., 1995). To the extent that both
antisocial personality and alcoholism cO-occur in parents, children in these homes may be
particularly vulnerable to both antisocial behavior and substance abuse themselves.

The estion of Multi le Risk

Zucker (1976) has suggested that, as the primary socialization factors in a young child’s
life, "the parent reward structure and modeling alternatives available within the family for
imitation" (p. 226) may foster a deviant pattern of antisocial behavior in the child. This, in turn,
disrupts the family’s affectional relationships and contributes to tension and conﬂict in the home.
The literature reviewed above points to parental alcoholism, antisocial behavior, and depression,
and to marital conﬂict as correlates of child-rearing practices and child antisocial behavior. Other
studies (e.g. Richman, et al. , 1982; McCord, 1979) indicate that these characteristics may Operate

as multiple risk factors for the development of child antisocial behavior, and later substance

25
abuse. Richman (et al., 1982) found that parent psychological distress, marital conﬂict, and

child—directed parent hostility were all related to child behavior. Such multiple indicators of
disharmonious relationships in the family were also associated with the later development of
antisocial behavior problems among children who had not exhibited these difﬁculties earlier.
Similarly, McCord’s (1979) study of boys in a program designed to prevent delinquency revealed
the impact of multiple family risk factors both on the development of antisocial behavior and on
the effectiveness of a treatment program intended to prevent such an outcome. Several factors
relating to the boys’ home environment were associated with their criminal behavior as adults:
high parental conﬂict and aggression were both positively related to crimes against persons
committed by the sons, while lower maternal affection and greater father deviance (alcoholism
and/or criminality) were positively related to property crimes. Level of supervision and mother’s
self-conﬁdence were related to both types of antisocial activity. Given the developmental
continuity between child antisocial behavior and later alcoholism, this literature suggests that both
of these aspects of child functioning are inﬂuenced by multiple risk factors.
The Prevention of Alcoholism

Hawkins and his colleagues (Hawkins, et al., 1992) have proposed that efforts to prevent
the abuse of alcohol and other substances must adopt a risk-focused approach, in which attempts
are made to modify and/or eliminate precursive factors. In light of evidence that the risk for
substance abuse is increased in the presence of multiple risk factors, preventative efforts should
address more than one area of difﬁculty. It is also recommended that preventative interventions
target populations that are at the greatest risk for developing problems with alcohol. Similarly,
others (e. g. Weber, et al., 1989) have emphasized the need to tailor intervention strategies to
speciﬁc risk factors within each targeted population.

Given the evidence that parent alcoholism, poor parenting practices, marital conﬂict, and

26

child antisocial behavior are all precursive to the development of alcoholism (Hawkins, et al.,
1992), prevention programs should target child populations in which several of these risk factors
are prevalent. Other prevention programs appear to operate under the assumption that all
individuals are at relatively equal risk for developing problems with alcohol, "an assumption that
is clearly at variance with the evidence" (Alterman & Tarter, 1983, p. 153). A better strategy
might be to identify and target populations that are at heightened risk for later alcoholism. In
light of substantial documentation, described above, that children who develop problems with
antisocial behavior at a young age are at risk for substance abuse later on, and that sons of
alcoholics are highly susceptible to both early aggressive behaviors and later alcoholism, young
male children in alcoholic homes represent a critical group for preventative efforts. Further, the
associations found between parent and child antisocial behavior in general (e. g. Chassin, et al.,
1992), and antisocial alcoholism and childhood risk in particular (e. g. Zucker, et al., in press),
indicate that the population of antisocial alcoholics and their children is one especially appropriate
for early intervention.

Such an approach is further supported by ongoing research, which has identiﬁed a variety
of differences between alcoholic and community contrast families participating in the Michigan
State Longitudinal Study (Zucker, Noll, & Fitzgerald, 1988). Focusing on the development of
preschool-age sons of alcoholic fathers, a substantial proportion of whom meet formal diagnostic
criteria for antisocial personality (Zucker, Noll, & Fitzgerald, 1988), this research has already
found that a) the high-risk male children are more familiar with alcoholic beverages than are the
matched control children (Noll & Zucker, 1983), b) there is a signiﬁcant relationship between
parental alcoholism and intrafamilial conﬂict and violence (Reider, et al. , 1988, 1989), c) parent-
child relationships are compromised by depression, antisocial behavior, and alcohol problems

among the parents (Davies, et al., 1989), and (1) these young sons of alcoholic fathers tend to

 

 

——...__.._.:-——_

27

engage in more aggressive and antisocial behaviors than their community counterparts (Fitzgerald,
et. al., 1988).

The prevalence of these multiple risk factors within this group of alcoholic families
supports efforts to target this population for preventative efforts. Since sons of antisocial
alcoholics are at particularly high risk for conduct disorder and its sequelae in childhood, and for
alcoholism and antisocial personality disorder later on, the impact of interventions on the
subsequent functioning of male children should represent an important marker of the effectiveness
of alcoholism prevention programs. With these considerations in mind, the Michigan State
University Multiple Risk Child Outreach Program (Maguin, Zucker, & Fitzgerald, 1994; Zucker
& Noll, 1987) was conducted as a prevention program designed to reduce the onset and/or
progression of antisocial behavior problems among preschool—age sons of alcoholic fathers, by
modifying child management practices used by the parents. Because of the continuity between
early antisocial behavior and later alcoholism, it was anticipated that addressing child behavior
problems early on would ultimately have a preventative effect on later displays of both antisocial
behavior and substance abuse. However, given the evidence that these are multiproblem families,
the program also sought to improve marital communication and conﬂict-resolution skills, and to
begin to address problems with alcohol abuse within the family. The longer-term goal was to
reduce the prevalence of alcoholism among the male target children, by addressing these multiple
risk factors early on.

The remainder of this chapter brieﬂy highlights the empirical support for the particular
treatment approach used in this study, provides a description of the strategies and theoretical
rationale upon which the Michigan State program is based, and discusses factors that have been
identiﬁed as determinants and disrupters in the treatment of child antisocial behavior. The ﬁnal

segment of the literature review examines current parent training process and outcome research,

28

and describes pretreatment parent and family characteristics that have been found to inﬂuence
treatment investment or treatment success.
Parent Training Approaches to Child Antisocial Behavior

Although several types of treatment approaches for antisocial children have been
developed, parent skills training strategies have received extensive empirical support as an
effective means of intervening in families with an antisocial child (e. g. Dumas, 1989; Gard &
Berry, 1986; Kazdin, 1985, 1987). The effectiveness of parent training approaches is consistent
with the conduct disorders literature, which has identiﬁed parenting variables related to harsh,
inconsistent discipline and poor supervision (Loeber & Dishion, 1983; Loeber & Stouthamer-
Loeber, 1986) as the most reliable predictors of future antisocial behavior. Common among the
various forms of parent training is the focus on parent-child interactions in the home. Of
particular concern are the coercive exchanges that Patterson (e.g. 1986) and others (Dadds, 1987;
Wahler & Dumas, 1987) have identiﬁed as playing a central role in promoting aggressive child
behavior. Parents are ﬁrst taught to identify, deﬁne, and observe problem behaviors in a new
way, and then learn procedures to implement at home. The therapist’s role is to instruct and
guide parents in the use of positive reinforcement, selective attention, mild punishment,
negotiation, and contingency contracting techniques to modify those target behaviors. The
sessions enable the parents to learn and practice these new techniques and to discuss their
implementation in the home. The goal, then, is to develop speciﬁc child-rearing skills in the
parents that will enable them to effectively monitor and manage their children and to break out
of the hostile exchanges, coercive cycles, and inconsistent patterns that characterize parent
interactions with antisocial children (Jaffe, et al., 1986; Jouriles, Barling, and O’Leary, 1987;
Home, 1981; Kratcoski, 1982; Olweus, 1980; Richman, et al., 1982; Stewart & Leone, 1978;

Walker, Downey, & Bergman, 1989).

29

One such program, Social Learning Therapy, provided the framework for the current
intervention. This approach is based on Gerald Patterson’s work in modifying coercive family
systems thought to contribute to child antisocial behavior (Patterson, 1976; Patterson,
Chamberlain, & Reid, 1982). Social Learning Therapy is somewhat different from more typical
parent training interventions in that it incorporates speciﬁc instruction for parents regarding social
learning principles. Thus, parents are not only taught how to modify speciﬁc behavior problems,
but are also educated about the underlying rationale governing the particular parenting techniques.
Those who have included this component into parent training interventions have argued that
providing parents with speciﬁc training in social learning principles is more efﬁcient than
focusing solely on discrete target behaviors (Patterson, Cobb, & Ray, 1973), that parents need
to understand the theoretical framework in order to adequately employ the techniques (Tharp &
Wetzel, 1969), and that generalization to other parent-child interactions is more likely to occur
when parents understand the principles behind the interventions (Forehand & Atkeson, 1977).
There is also some empirical evidence that treatment-related improvements in child behavior and
parenting are enhanced when parents are instructed about social learning principles in addition
to learning how to manage speciﬁc problem behaviors (McMahon, Forehand, & Griest, 1981).
The speciﬁcs of the program used in the current intervention will be described in further detail
later on.

Parent Training and Substance Abuse Prevention

Given the well-documented association between child antisocial behavior and the
development of substance abuse problems (Kandel, 1982; Loeber, 1982; Robins, 1966), as well
as the role that parent supervision and monitoring behaviors appear to play in both child antisocial
behavior (Loeber & Dishion, 1983; Loeber & Stouthamer-Loeber, 1986) and adolescent substance

abuse (e. g. Chassin, et al., 1993), parent skills training and family therapy programs have been

30

identiﬁed as effective approaches in the prevention of alcohol and other drug problems (Alvy,
1991; Hawkins, et al., 1992; McKay, 1991; Popkin, 1991). Although most of the research on
such efforts has focused on the ability of these programs to reduce antisocial behavior problems
(e.g. Fleischman, 1981; Patterson & Reid, 1973; Peed, Roberts, & Forehand, 1977) and to
prevent or reduce juvenile delinquency (e.g. Alexander & Parsons, 1973; Klein, Alexander, &
Parsons, 1977; Tremblay, McCord, Boileau, LeBlanc, Gagnon, Charlebois, & Larivee, 1990),
the success of this approach in improving family management skills, reducing immediate problem
behaviors, and preventing subsequent antisocial acts supports its continued use as an intervention
program. However, it should also be noted that subjects in such studies are typically selected
because of problem child behavior, rather than because of particular characteristics of the parents.
Although little is known thus far about the extent to which parent training interventions are
successful in preventing substance abuse among children with alcoholic parents, there is some
direct evidence that parenting skills training with substance abusing parents is successful in
reducing child behavior problems, improving parenting practices, and reducing children’s interest
in using alcohol and cigarettes (DeMarsh & Kumpfer, 1986). Similarly, a parent training
intervention with parents of high-risk youth indicated a positive trend in the prevention of tobacco
use among the target children (Dishion, Kavanaugh, & Reid, 1989).
Parent and Famin Disrupters in the Treatment Of Child Antisocial Behavior

The role of marital conﬂict and parent psychopathology in disrupting child rearing
practices and contributing to the developement of child antisocial behavior (e. g. Davies, et al. ,
1989; Emery & O’Leary, 1982; Johnson & Lobitz, 1974; Johnson & O’Leary, 1987; Lahey,
et al., 1988) suggests that these family characteristics may also inﬂuence the success of
interventions targeting parenting practices and child behavior. The following section reviews the

small body of literature to date that has examined the effects of these potential disrupters on

31

family-based parent skills training interventions.

Marital conﬂict. As described above, several studies have found an association between
marital conﬂict and parenting behavior (Gottman & Katz, 1989; Johnson & Lobitz, 1974; Jouriles
& Farris, 1992; Kemper & Reichler, 1976). One explanation for this relationship has been
offered by Emery (Emery, Hetherington, & Dilalla, 1984), who suggests that parenting is often
compromised when interspousal conﬂict and marital dissatisfaction are high. From this
perspective, marital conﬂict promotes inconsistencies in parenting, such that mothers and fathers
may employ different practices with the same child, and each parent may him/herself respond in
inconsistent ways from one time to the next. Such inconsistent parenting has been associated with
negative child outcomes in a number of studies (Block, Block, & Gjerde, 1986; Block, Block,
& Morrison, 1981; Chassin, et al., 1993; Emery, et al., 1984; Gottman & Katz, 1989;
Hetherington, Cox, & Cox, 1981; Patterson, 1980).

While marital conﬂict has been found to affect daily child-rearing practices and child
behavior, the extent to which this factor also disrupts the treatment of child antisocial behavior
is less clear. Oltrnanns and his colleagues (Oltrnanns, Broderick, & O’Leary, 1977) found a
negative correlation between marital adjustment and the initial severity of children’s behavior
problems among a clinic sample. However, pretreatment level of marital discord was not
predictive of child behavior change either at termination or at a 5-month follow-up. In contrast,
Dadds and his colleagues (Dadds, Sanders, Behrens, & James, 1987 ; Dadds, Schwartz, &
Sanders, 1987) found that the extent of marital discord reported by parents of conduct-disordered
children was related to parent-to-child behavior at a six-month follow-up, with greater marital
discord associated with more impaired parenting. However, marital conﬂict was not related to
immediate post-treatment measures of parenting behavior. Thus, marital discord impeded the

later maintenance of treatment effects, even though it did not affect initial progress or gains.

32

In earlier research with the sample of families used in the present study, interspousal
aggression and conﬂict were not signiﬁcantly associated with child outcome either at mid-
treatment or at the end of the intervention (Nye, 1992). Although it is possible that a post-
treatment follow-up might have revealed such an effect, an alternative explanation is that the
potential impact of marital conﬂict on child outcome was reduced through the intervention, as
parents who completed the program were engaged in marital problem-solving work in addition
to the child-focused component. In support of this explanation, other research has indicated that
parent training interventions incorporating partner enhancement strategies are more effective both
in changing deviant child behavior and in maintaining those effects than is parent training alone
(e. g. Griest, Forehand, Rogers, Breiner, Furey, & Williams, 1982). In fact, in the Dadds studies
(Dadds, et al. , 1987a; 1987b), marital conﬂict was differentially predictive of child post-treatment
outcome depending on the nature of the intervention. Distressed couples who received partner
support training in addition to the child management component were as able to maintain
treatment gains and positive parenting behaviors as were couples in nondiscordant marriages,
whereas this effect was not found for distressed couples who received only child management
training. Thus, marital conﬂict has not been found to signiﬁcantly impair treatment success in
programs that explicitly address marital difﬁculties within the intervention protocol.

Parent psychopathology. Although there is a paucity of research in this area, the
literature also indicates that pretreatment levels of parent psychopathology are predictive of both
treatment participation and outcome in parent training programs. Maternal depression, for
example, has been associated with treatment attrition rates (McMahon, Forehand, Griest, &
Wells, 1981b), and with parents’ failure to participate in treatment outcome evaluations as follow-
up (Griest, Forehand, & Wells, 1981). Similarly, Furey and Basili (1988) found higher levels

of depression among mothers who dropped out of treatment than among those who continued with

33

the intervention protocol. Thus far, alcoholism has not been studied as a predictor of
participation, although the literature frequently has indirectly linked this problem given the strong
familial assortative relationships between alcoholic men and depressed women (cf. Merikangas,
et al., 1985).

Regarding treatment outcome, Dumas and Wahler (1983) found that level of positive
improvement in parenting practices at the end of treatment was impaired by higher levels of
maternal psychopathology and marital violence, and by lower family socioeconomic status. A
later study by Dumas (1986), however, suggests that the primary inﬂuence of parent
psychopathology on treatment outcome may be through its association with low socioeconomic
status. Similarly, Webster-Stratton (1985) found socioeconomic disadvantage to be highly
predictive of more negative child outcomes in parent training for conduct-disordered children.
Pretreatment depression among the mothers was a much weaker predictor, although it was
strongly negatively related to socioeconomic status. Overall, then, the role that various types of
parent difﬁculty play in reducing the ability of parent training programs to increase parenting
skills and reduce child antisocial behavior has been unclear. The literature has focused largely
on parent depression, with much less attention given to parent alcoholism. Further, participants
in all of these studies were identiﬁed because of problem child behavior, rather than for
characteristics of the parents. Thus, it is difﬁcult to generalize ﬁndings from this research to our
own sample of families, who were explicitly recruited for the intervention program because of
paternal alcoholism.

The literature does indicate, however, that both child-rearing practices and childantisocial
behavior are differentially impaired by parent depression, antisocial behavior, and alcoholism,
and by family factors related to marital conﬂict and socioeconomic deprivation. Further, some

of these factors have been found to reduce the effectiveness of parent training programs aimed

34

at changing child-rearing strategies, particularly in treatment protocols which do not address
parent and family difﬁculties that extend beyond problem child behaviors. However, there is
evidence that processes relating to the course of the intervention itself may account for or
moderate the effects of parent and family characteristics on treatment outcome.
Treatment Investment and Treatment Outcome

Patterson and Chamberlain (1988) have proposed that parent and family trouble inﬂuences
the extent to which parents commit to and cooperate in parent training programs. From this
perspective, it is the variable of parent cooperation or compliance that is most directly related to
treatment outcome, through its association with change in parenting practices. In fact, Patterson
and his colleagues (e. g. Patterson, Chamberlain, & Reid, 1982) have demonstrated that higher
levels of marital conﬂict and parent psychopathology and fewer family resources interfere with
parent training by increasing client resistance to the treatment and disrupting the acquisition and
application of appropriate child rearing practices.

Treatment investment has been a distinct focus of psychotherapy research (e.g. Nelson
& Borkovec, 1989; Orlinsky, 1989), and compliance has long been considered requisite for
positive change in behavioral and cognitive-behavioral approaches (Burns & Nolen-Hoeksema,
1991; Shelton & Levy, 1981). For example, in corroboration of other authors (e. g. Kirtner &
Cartwright, 1958; Rice & Wagstaff, 1967; Saltzrnan, Luetgert, Roth, Creaser, & Howard, 1976),
Gomes-Schwartz (1978) identiﬁed patient involvement with and engagement in therapy as the best
predictor of outcome in individual psychotherapy. Yet, this treatment process dimension has
rarely been considered in outcome research involving behavioral types of family interventions
(Primakoff, Epstein, & Covi, 1986). As Primakoff and her colleagues (1986) point out,
treatment outcome research that fails to examine treatment investment is likely to be inﬂuenced

by "compliance bias" (Feinstein, 1979), whereby purported differences among various treatment

35

groups may be more related to the extent of patient compliance with the regimen than to the
impact of the program itself. Thus far, only a few studies on family-based interventions have
examined the inﬂuence of parent treatment compliance or investment on treatment outcome.

In their examination of factors affecting outcome in a parent training program for
mentally retarded children, Clark and Baker (1982) found that pretreatment levels of family
socioeconomic status, income, and maternal education were predictive of parent skill acquisition
by the end of the program, with higher levels of these family resources predicting greater parent
proﬁciency. However, family resources were not associated with the extent to which parents
followed through with acquired skills at follow-up 14 months later. Rather, it was parents’
application of and compliance with newly acquired skills during the course of treatment that
predicted their continued proﬁciency over a year later. Thus, parents who were highly
cooperative with the program and who complied with homework assignments during treatment
were able to maintain treatment gains over time.

Earlier research with alcoholic families from the Michigan State University Multiple Risk
Child Outreach Program examined the role of parent psychopathology, marital conﬂict, family
socioeconomic deprivation, and maternal treatment investment in predicting child outcome (Nye,
et al., in press). Because of colinearity among the various measures of parent pathology, the
psychopathology variable used was a composite of lifetime alcohol problems, current and worst-
ever depression, and antisocial behavior. For the same reason, investment was a composite of
cooperation with homework between sessions, and engagement in and openness to the therapeutic
work within treatment sessions. Initial stepwise regression analyses of pretreatment parent and
family characteristics identiﬁed father psychopathology as a signiﬁcant positive predictor of
negative child behavior at a mid-treatment time point. However, maternal treatment investment

accounted for a signiﬁcant additional portion of the variance beyond that accounted for by

36

paternal distress. Similarly, although higher maternal psychopathology was predictive of less
positive child behavior at both mid-treatment and the end of the program, maternal treatment
investment was also a signiﬁcant positive predictor. Further, analyses indicated that the inﬂuence
of maternal psychopathology on child outcome was accounted for by the association between
pretreatment maternal distress and subsequent treannent investment. In this regard, mothers with
higher levels of personal trouble appeared to have greater difﬁculty cooperating with and
engaging in the treatment regimen, which in turn impeded treatment progress (Nye, 1992).
Finally, whereas both family resources (income, parent education, and socioeconomic status) and
maternal treatment investment were negatively predictive of negative behavior at the end of the
intervention, maternal treatment investment was found to moderate the deleterious impact of
economic deprivation on treatment success. That is, families in which the mothers were highly
invested in and cooperative with the program showed a reduction in negative child behavior
regardless of their economic resources. In contrast, lack of family resources predicted greater
negative child behavior at the end of the program among families in which the mother was not
highly invested in the success of the regimen. Together, these ﬁndings indicate that treatment
investment represents a more proximal determinant of treatment success and positive child
outcome than is pretreatment level of parent psychopathology and family economic deprivation.
The above results are consistent with research by Patterson and Chamberlain (1988), who
also used measures of both within-session resistance to therapists’ instructions, and between-
session noncompliance with homework assignments as markers of treatment investment among
parents of preadolescent antisocial children. Their preliminary ﬁndings suggested that family
stressors such as marital conﬂict, parent psychopathology, and low socioeconomic status
indirectly impeded positive child outcome by increasing within-session conﬂict and decreasing

treatment compliance. Subsequent analyses conﬁrmed that social disadvantage, parent depression,

37

and parent antisociality inﬂuenced level of resistance in and cooperation with treatment (Patterson
& Chamberlain, 1994). Further, these researchers found that decreased resistance during parent
training was associated with improved discipline, problem solving, and monitoring practices by
the parents (Patterson & Stoolmiller, in press).

These studies indicate, then, that positive change in targeted parent and child behavior
arising through parent training interventions are more proximally accounted for by parent
treatment investment, and more distally by parent psychopathology and family economic
deprivation. Further, the research suggests that parent treatment investment may serve to
promote treatment gains that might otherwise be impaired by parent and family difﬁculty.
Clearly, then, a more detailed examination of treatment process characteristics, including parent
treatment investment, is a necessary next step in determining more explicitly what factors
contribute to successful outcomes in parent training approaches to prevent child antisocial
behavior and substance abuse.

Predictors of Treatment Investment: Dropout Studies

Several studies have explored factors associated with families that prematurely disengage
from parent training programs and/or discontinue their involvement with post-treatment
evaluations. In their examination of pretreatment variables predicting participation in follow-up
assessments in a program for mothers of noncompliant children, Griest and colleagues (Griest,
et al., 1981) found that maternal depression, but not marital satisfaction, was negatively related
to this measure of cooperation. Similarly, Griest and Forehand (1982) have associated parental
depression with an increased dropout rate in parent training. Furey and Basili (1988) found that
both depression and socioeconomic disadvantage were predictive of mothers who dropped out Of
such a program, and that these mothers also had more impaired parenting skills at the beginning

of treatment than did mothers who completed the program. In addition, mothers who remained

38

in the program but did not perceive a positive improvement in their child’s behavior had
signiﬁcantly higher levels of depression than mothers who were satisﬁed with the treatment.
Other research has also identiﬁed low socioeconomic status, as well as less parent education, as
predictors of early termination from parent training (Dumas & Albin, 1986; McAuley, 1982;
McMahon, et al., 1981b). Again, families involved in these studies were all identiﬁed by way
of their child’s antisocial behavior.

Among alcoholic families involved in the Michigan State University Multiple Risk Child
Outreach Program, no differences were found at baseline in level of child behavior problems,
parent psychopathology, marital conﬂict, or family resources between parents who completed
treatment and those who terminated prematurely (Nye, 1992). However, a few differences were
noted between families that participated in treatment and/or assessments throughout the course
of the study, and those who either refused to participate or became ineligible for the program
because of divorce or relocation. Speciﬁcally, mothers in families that disengaged from the
larger research project reported more extreme rates of certain types of marital violence, and
higher current levels of maternal alcohol consumption. To some extent, then, the more
dysfunctional families may have been selected out of the project, via marital separation and/or
refusal to participate in follow-up assessments.

Interestingly, some research with alcoholics has indicated that patients with lp_\y_e_r_ levels
of depression at intake are not as likely to persist in treatment as are alcoholics who are more
depressed (MacMurray, Nessman, Haviland, & Anderson, 1987). The authors propose that a
certain level of distress is necessary in order for patients to be motivated to continue with
treatment. In an extensive review of earlier research on treatment dropouts, Baekeland &
Lundwall (1975) identiﬁed three clusters of characteristics associated with premature termination:

1) intrapsychic factors of the patient that either favorably or unfavorably dispose him/her to

39

psychological intervention; 2) factors related to the personality, style, and attitudes of the
therapist; and, 3) environmental factors which may interfere with continued participation in
treatment.

Although the treatment dropout literature represents an important step in identifying
personal and interpersonal characteristics that are associated with treatment participation, dropout
and refusal rates are only crude indices of treatment investment and compliance and are not
necessarily indicative of the treatment process. There is growing evidence that clients who
terminate treatment prematurely do not represent a homogeneous group (Baekeland & Lundwall,
1975; Fiester & Rudestam, 1975; Pekarik, 1992; Schwartz, 1991). For example, Pekarik (1992)
found that some parents who had sought assistance in managing problem child behaviors
terminated prematurely because of dissatisfaction with treatment or because of environmental
obstacles impeding treatment attendance, while other parents terminated because the target child’s
behavior had improved. Thus, it is not feasible to assume that all drop-outs are less invested in
treatment than are those who follow the designated timeline. Nor is it appropriate to conclude
that those who terminate early are treatment " failures ". In the Pekarik (1992) study, children
whose parents terminated treatment prematurely because of problem improvement exhibited
comparable treatment outcome to children whose parents continued in treatment. It is also faulty
to assume that clients who remain in treatment will uniformly experience a positive outcome
compared with treatment drop-outs. In fact, in earlier work with the current treatment sample,
only highly invested mothers completing the parent training program were able to effect positive
changes in their child’s behavior at the end of treatment. In contrast, neither the less invested
treatment completers nor the mothers who dropped out of the program experienced signiﬁcant
child behavior change (Nye, Zucker, & Fitzgerald, in press). Thus, although premature

termination may be one index of level of treatment investment, this variable alone does not

40

provide an understanding of the processes that either foster or impair treatment success.
Predictors of Treatment Investmen_t: Treatment Process Studies

Using a more process-related measure of investment, Dumas (1986) examined parent and
family characteristics related to combined measures of treatment involvement and treatment
outcome. Treatment outcome, as determined by positive change in parenting skills, and treatment
involvement, as measured by attendance at scheduled meetings and compliance with program
instructions, were most directly impaired by lower family socioeconomic characteristics.
Maternal psychopathology and marital violence were more indirectly related to treatment
involvement and outcome, through their association with characteristics of the socioeconomic
setting.

Patterson and his colleagues have undertaken a series of studies on both within-session
resistance to therapists’ interventions and between-session noncompliance with homework
assignments as measures of treatment investment. In general, the authors have reported a high
rate of overt opposition to the therapist and low compliance with homework among parents with
criminal and/or antisocial histories (Patterson & Chamberlain, 1992) and, to a somewhat lesser
extent, by depressed parents (Chamberlain & Ray, 1988). Typical of most of the parent training
research, these studies have focused on parents of antisocial or conduct—disordered children, with
one recent study targeting a clinical sample of 70 preadolescent boys and girls who were referred
for treatment because of extreme antisocial behavior problems (Patterson & Chamberlain, 1994).
Most of the target children were boys, with a mean age of 9 years. Participants were
economically disadvantaged, and less than half were two—parent families. Analyses indicated that
both maternal pathology (depression, antisociality, and stress) and social disadvantage were
associated with initial levels of mothers’ resistance, but were not signiﬁcantly related to maternal

investment at the end of the intervention. In contrast, paternal antisociality, depression, and

41

social disadvantage were associated with resistance and reluctance throughout treatment (Patterson
& Chamberlain, .1994). Thus, there is some suggestion that individual and social stressors may
differentially inﬂuence mother and father investment in the course of treatment.

Research on the young alcoholic families in the Michigan State program (Nye, 1992)
examined the extent to which pretreatment levels of child behavior problems, mother and father
psychopathology, marital conﬂict, and family socioeconomic deprivation served as predictors of
maternal treatment investment through mid-treatment and through termination. Only the maternal
psychopathology cluster, comprised of current and worst-ever depression, antisocial history, and
lifetime alcohol problems, was Signiﬁcantly associated with the extent to which mothers
demonstrated investment in the treatment program at both time points. Additional work is
necessary to examine differences in these relationships from pretreatment to mid-treatment in
comparison with the mid-treatment to termination phase. In addition, father treatment investment
needs to also be considered as a treatment process variable.

Other Treatment Process Characteristics

Based on the literature reviewed thus far, it is evident that the effectiveness of parent
training and other family-focused interventions targeting aversive child behavior is strongly
associated with the extent to which parents comply with homework assignments, are engaged and
cooperative with the therapist, and are invested in the success of the protocol. Further, the
importance of treatment investment to treatment success is maintained even when the inﬂuence
of more distal parent and family characteristics, such as parent psychopathology, marital conﬂict,
and family socioeconomic status, is taken into account. In addition, there is some evidence that
treatment investment is not only a vehicle through which some of these other factors affect
outcome, but also a potential buffer from the negative inﬂuence that family difﬁculty can have

on treatment SUCCESS .

42

Treatment investment, as reﬂected in parent Openness to change and cooperation with
assignments, is only one aspect of the treatment process that is likely to be associated with
treatment outcome. Other characteristics, such as parents’ expectations about a program and their
satisfaction with its results during treatment, are likely to be related to both investment and
outcome. Similarly, therapists’ expectations about a family’s ability to beneﬁt from a treatment
protocol may inﬂuence the course and ultimate success of the intervention. The remainder of this
literature review describes research which has examined the inﬂuence of treatment expectations
and satisfaction on investment and outcome.

Treatment satisfaction. The treatment process research identiﬁes treatment satisfaction
as another important treatment characteristic. McNeill and his colleagues (McNeill, May, & Lee,
1987) reported that their sample of college students who terminated treatment prematurely were
less satisﬁed with the services they received and viewed their counselors more negatively than
did subjects who terminated "successfully". Similarly, a subsample of parents in the Pekarik
(1992) study who terminated their involvement in child-focused therapy did so because Of
dissatisfaction with treatment services. However, the relationship between treatment satisfaction
and treatment investment is not perfect, at least when investment is operationalized as
continuation in therapy. As noted above, several studies have found adequate levels of
satisfaction among treatment drop-outs (e. g. Heineman & Yudin, 1974; Kline, Adrian, & Spevak,
1974; Pekarik, 1992). For example, in Heineman and Yudin’s (1974) study, 50% of treatment
drop-outs reported satisfaction, while Kline and colleagues (Kline, et. al., 1974) found a
satisfaction rate of 59% among their group of drop-outs. This relatively high level of satisfaction
among clients who drop-out of treatment provides further support for the use of measures of
treatment investment that extend beyond mere attendance at sessions.

There is also some evidence that client satisfaction with treatment is associated with

43

expectations for continued improvement (Hansson & Berglund, 1987) which may, in turn, affect
clients’ investment in maintaining treatment gains. For example, Furey and Basili (1988)
examined factors related to mothers’ perceptions of improvement in their children’s noncompliant
behavior. Mothers who reported an improvement in their child’s behavior from pretest to the
two-month post-treatment follow-up were classiﬁed as satisﬁed with the intervention, while
mothers who continued to report child maladjustment at follow-up were deﬁned as dissatisﬁed.
Observations of parent and child interactions indicated comparable levels of improvement in both
child noncompliance and parenting behavior. Thus, despite objective improvements as a result
of the intervention, "dissatisﬁed" mothers perceived a lack of positive change in child behavior
at the two-month follow-up. This represented a deterioration from immediate end-of-treatment
levels, as mothers in the dissatisﬁed group reported initial treatment gains from pretest to
termination. An examination of pre-treatment levels of parental distress indicated that mothers
in the dissatisﬁed group were signiﬁcantly more depressed than those in the satisﬁed group,
which may account for the deterioration in perceptions of improvement among dissatisﬁed
mothers. The authors conclude that:

many mothers do not maintain positive perceived gains in their child’s adjustment

after treatment despite continued maintenance of the parenting skills learned in

therapy and in the child’s observed behavior. The failure to establish enduring

improvements in maternal perceptions creates the risk that the mothers will

discontinue the behavioral management techniques which, in fact, have reduced

deviant child behavior. . .Hence, more than the child’s actual behavior inﬂuences

a mother’s perception of her child and, ultimately, continued compliance with

treatment (pp. 562-563).

Although limited by its failure to obtain direct measures of satisfaction from the
participants, this study highlights the potential inﬂuence that client perceptions of treatment gains
may have on continued investment. However, an additional limitation of this study is its failure

to consider levels of satisfaction during the course of treatment. Much of the research

incorporating satisfaction measures has looked only at client satisfaction at termination (Furey &

44
Basili, 1988; Hansson & Berglund, 1987; Hobbs, Walle, & Hammersly, 1990). Such studies do

not permit an examination of the inﬂuence of client satisfaction Mg treatment on subsequent
treatment investment, treatment outcome, and expectations of continued treatment success.
Clearly, such research is necessary in order to delineate the role of client satisfaction in enhancing
or disrupting the treatment process.

The link between satisfaction and outcome is also unclear. Although Pekarik (1992)
found poor outcomes for dissatisﬁed adult and child dropouts, ongoing adult clients who reported
very high satisfaction ratings also reported low symptom improvement. Edwards and colleagues
(Edwards, Yarvis, Mueller, & Langsley, 1978) found variations in client satisfaction throughout
treatment among outpatient adults. The authors noted that satisfaction increased from the second
visit to the last visit, but did not show such a trend from session two through follow-up. In
addition, the relationship between satisfaction and outcome differed at these varying time points;
client satisfaction was most poorly correlated with improvement at the second visit and most
strongly associated with outcome at the last visit. Thus, level of satisfaction at one point in
treatment may have different implications for outcome than at other time points.

Research conducted by Tracey (1989) indicates that treatment satisfaction may progress
in a curvilinear trend from the beginning of treatment to the end. In this study, both clients and
therapists reported high satisfaction early in treatment, lower satisfaction at mid-treatment, and
a return to high satisfaction by termination. However, this trend was found only for clients who
had experienced the most improvement by the end of treatment. In contrast, the less successful
outcome groups demonstrated a greater level of continuity in client and therapist satisfaction over
time. The author proposes that difﬁcult issues are typically addressed during the middle sessions
of successful interventions, thereby creating less client and therapist satisfaction with the process

than during earlier and later stages. Under this formulation, stability in satisfaction during the

45
course of therapy may be a marker that critical issues are not being addressed, thereby impeding
the ultimate success of the intervention.

Lebow (1982), in an extensive review of client satisfaction research, summarizes the
literature in this way:

The research suggests that satisfaction is closely related to...clients’ global

assessments of their success in treatment; low-to-moderately related to therapists’

satisfaction with treatment, to clients’ more Speciﬁc view of change, and to the
tendency to terminate prematurely; and only slightly related to therapists’

assessments of change (p. 253).

Thus, multiple measures of satisfaction and outcome are necessary in order to understand the
complex processes taking place in the context of treatment.

Treatment expectations. Baekeland and Lundwall’s (1975) review of earlier research on
premature disengagement from treatment identiﬁed a variety of client and therapist characteristics
associated with this index of investment. Among these were the initial motivation of the patient
to engage in treatment, therapist expectations of patient improvement, and discrepant treatment
expectations between client and therapist. In terms of the latter, all 6 studies reviewed indicated
that the match between patient and therapist expectations played an important role in continued
participation in individual psychotherapy. Although not representing an extensive body of
research, this early literature supports the role of treatment expectations in treatment investment,
as measured by continuation in therapy. Other researchers (e.g. Kazdin, 1981; Wolf, 1978)
have proposed that a client’s early attitudes about a treatment program will be Signiﬁcantly
associated with his/her investment in and cooperation with the intervention procedures.
Similarly, Bandura (1977) has argued that initial expectations about the usefulness of treatment
inﬂuence clients’ motivation and investment, as well as their willingness to continue with

treatment when faced with challenges or difﬁculties. Thus far, however, there have been few

empirical investigations of the relationship between expectations and ongoing investment.

46

Regarding treatment success, a more substantial body of literature has identiﬁed a
relationship between client expectations and outcome with a variety of patient populations. These
studies indicate a positive relationship between favorable pretreatment expectancies and eventual
change in the targeted problem (Bloch, Bond, Qualls, Yalom, & Zimmerman, 1976; Bowden,
Schoenfeld, & Adams, 1980; Collins & Hyer, 1986; Duckro, 1991; Evans, Kiolet, & Smith,
1985; Evans, Smith, Halar, & Kiolet, 1985; Waas & Anderson, 1991; Wilkins, 1973). In one
of the few studies focused on parent training interventions, Clark and Baker (1983) examined
factors inﬂuencing parent proﬁciency in teaching skills and managing problem behaviors with
their mentally retarded children. Parents who were less proﬁcient in applying new skills at the
end of the program scored higher on pretreatment expectations of obstacles to teaching their
children than did the more proﬁcient parents. Further, this difference in expectations was not
associated with differences in pretest child characteristics between the two groups.

Elliott (1986) has suggested that a client’s pretreatment evaluation of an intervention will
directly inﬂuence his/her use of intervention procedures, thereby affecting treatment outcome,
as well. Overall, the literature supports the proposed relationships between pretreatment
expectations, ongoing treatment investment, and post-treatment outcome. However, the
interrelationships among these three treatment variables has yet to be examined. In addition, the
inﬂuence of the therapist on treatment process and treatment outcome characteristics needs to be
examined more closely. As described earlier, studies Of premature treatment termination have
identiﬁed the therapist as an important agent in the treatment process. In particular, the level of
agreement between client and therapist expectations has been found to predict whether or not
clients remain in therapy (Baekeland & Lundwall, 1975). In addition, there is some evidence that
the therapist’s initial expectations about a client’s potential to beneﬁt from treatment inﬂuence the

treatment process and subsequent treatment outcome (e.g. Duckro, 1991). Similarly, Hakan

47

(1990) proposes that therapist expectations at the outset of treatment affect client treatment
investment and dropout rates.

Wilkins (1973) has argued that the early research relating client expectancy to outcome
has been confounded by an exclusive reliance on patient self—reports of change, and proposes that
therapist expectations may account for expectancy effects that have typically been attributed to
clients. Some research has supported this position, and has identiﬁed therapists’ expectations to
be more predictive of client outcome than clients’ own expectations for symptom reduction
(Martin & Sterne, 1975). However, it must also be noted that this latter study was with a group
of severely disturbed patients, whose expectations were measured immediately upon
hospitalization. Such a patient sample might presumably have uniformly low expectations of their
own prognosis, given their extreme circumstances.

What may perhaps be a more crucial argument made by Wilkins (1973) is his caveat
against attributing a causal effect to client expectations, "when more directly observable
operations leading to improvement can be identiﬁed". From this perspective, expectations of
therapeutic gain are not in and of themselves sufﬁcient for treatment to be successful. Rather,
the processes involved in the treatment protocol must also be examined. In support of this,
research on a group of inpatient substance abusers found that patients with the most successful
outcomes were those whose expectations for success were initially m at intake, and who became
increasingly more optimistic about their prognosis as treatment progressed (Burling, Reilly,
Moltzen, & Ziff, 1989). Thus, the impact of initial client expectations on outcome, while
potentially important, may not be direct. Rather, it is likely that expectations are modiﬁed as a

function of the client’s experiences in treatment.

48

m

The literature on substance abuse in adolescence and early adulthood has identiﬁed a
variety of parent, family, and child characteristics that are thought to be early precursors to
alcoholism. Regarding child characteristics, for example, early antisocial behavior has been
associated with adolescent substance abuse (e.g. Robins, 1978; Johnston, et al., 1978; Kandel,
et al., 1978; Wechsler & Thum, 1973), and both Loeber (1986) and Zucker and Fitzgerald
(1991) have proposed developmental pathways linking child antisocial behavior with early
initiation into drinking, problem drinking, and progressively more severe antisociality. Despite
such continuity, however, there is also evidence that a substantial proportion of children who
display severe behavior problems do not manifest these problems later on (Ghodsian, Fogelman,
Lambert, & Tibbenharn, 1980; Loeber & Dishion, 1983; Robins, 1978). Thus, early child
antisocial behavior alone may not be an adequate marker of risk for later substance abuse. Such
co-occurring characteristics as parent alcoholism, poor family management practices, and family
conﬂict have also been identiﬁed as risk factors in the development of alcoholism. Further, there
is evidence that parental alcoholism has a particularly strong inﬂuence on adolescent substance
abuse, both via direct modeling (Chassin, et al., 1991; Kandel, 1982; N011 & Zucker, 1983) and
through its association with disrupted parenting (Chassin, et al., 1993; Davies, et al., 1989),
family conﬂict (Reider, et al., 1989), and child antisocial behavior (Fitzgerald, et al., 1989).

Given the conﬂuence of these early risk factors in alcoholic families, it is appropriate to
target this population with intervention efforts designed to minimize the risks and prevent the
subsequent development of alcoholism among the offspring of alcoholic parents. Although little
research to date has examined the longer-term impact of family-based prevention approaches on
youth substance use, parent skills training and family therapy programs have been found to be

effective in improving the poor parenting practices and negative child behavior thought to place

49
youth at risk for later problems with alcohol (e. g. Fleischman, 1981; Patterson & Reid, 1972).

Yet, these efforts have typically focused on antisocial children as the targeted high-risk group,
at the exclusion of particular parent and family characteristics. An alternative strategy is to
identify children whose early rearing environments expose them to a greater density of risk, and
who may therefore be most susceptible to the development of substance abuse problems later on.
This latter strategy may be particularly important given evidence that at least one form of
alcoholism, antisocial alcoholism, evolves from early childhood and involves a progressive
cumulation of risk that results in early and severe abuse of alcohol and other substances (Zucker,
in press). Further, the high prevalence of antisocial behavior and alcoholism in sons of alcoholic
fathers, and the unique contributions of paternal alcoholism and antisociality to adolescent alcohol
use and externalizing symptoms, respectively, suggests that young sons of antisocial alcoholic
fathers are exposed to a density of risk much greater than is otherwise found in the general
population.

The present study focuses on a community-based effort to target such a high-risk sample.
The Michigan State University Multiple Risk Child Outreach Program (Maguin, et al., 1994;
Nye, et al., in press; Zucker, et al., 1986; Zucker, et al., 1990) utilized an outreach protocol to
contact and treat a group of intact families with an alcoholic father and at least one preschool-age
male target child. This sample was expected to be particularly dense in child risk for later
alcohol and drug problems, by virtue of a high rate of antisociality and multiple parent and family
stressors (Zucker, et al., 1988). Given the support for family-based, parent skills training
programs both in the treatment of antisocial child behavior (Dumas, 1989; Gard & Berry, 1986;
Kazdin, 1985, 1987; Patterson & Chamberlain, 1992) and in the prevention of substance abuse
(Hawkins, et al. , 1992), the current program was designed to reduce the onset and/or progression

0f antisocial behavior problems among these preschool-age sons of alcoholic fathers, by

50

modifying child management practices used by the parents. In addition, the program sought to
improve marital communication and conﬂict-resolution skills, and to begin to address problems
with alcohol abuse within the family. The longer-term goal, then, is the prevention of alcoholism
among the male target children, by addressing these multiple risk factors early on.

In terms of the ﬁrst goal, that of addressing child behavior problems, a decrease in
antisocial behavior and an increase in prosocial behavior in the target children at the end of
treatment has already been observed; however, improvement in negative behavior was found to
have eroded by the six-month follow-up (Maguin, et al. , 1994). Subsequent research (Nye, 1992;
Nye, et al., 1993) identiﬁed variations in the extent to which families beneﬁtted from the
intervention. In corroboration of other parent training programs which included a marital
enhancement component (Dadds, et al., 1987a, 1987b; Griest, et al., 1982), pretreatment level
of marital discord was not associated with child outcome. Also consistent with the literature
described thus far (e.g. Dumas, 1986; Dumas & Wahler, 1983; Webster-Stratton, 1985), higher
levels of parent psychopathology and greater economic deprivation were associated with fewer
treatment gains among these families. However, it was also noted that maternal treatment
investment, as measured by compliance with homework assignments and cooperation with
therapeutic interventions, was a signiﬁcant predictor of child outcome, with children of mOre
highly invested mothers showing greater behavioral improvement. Further, it was determined
that maternal investment mediated the inﬂuence of maternal psychopathology on child behavior,
and moderated the deleterious effects of economic deprivation on child outcome.

Overall, then, research conducted thus far with families from the M.S.U. program is

Consistent with other studies (e. g. Patterson & Chamberlain, 1992) which have identiﬁed
treatment investment as a predictor of treatment outcome. However, other treatment process

Characteristics and treatment outcome measures need to be taken into account. For example, the

51

Patterson group has emphasized the importance of using multiple measures of child outcome, of
measuring change in parenting behavior as well as change in child behavior, and of including
measures of paternal treatment investment (e. g. Patterson & Chamberlain, 1992). In addition,
the literature reviewed thus far indicates that client and therapist self-report ratings of expectations
and satisfaction may be associated with treatment investment and treatment outcome. As a
pretreatment variable, patient and therapist expectations are thought to inﬂuence initial levels of
client engagement (Bandura, 1977; Kazdin, 1981; Wolf, 1978), as well as therapists’ attitudes
toward and/or experience of clients (Duckro, 1991). In addition, initial expectations have been
associated with eventual outcome in parent training interventions (Clark & Baker, 1983),
alcoholism treatment (Burling, et al., 1989), and other treatment approaches (e. g. Bloch, et al.,
1976; Evans, et al., 1985; Waas & Anderson, 1991).

As a process variable, the role of expectations remains largely unknown. Most of the
literature reviewed did not examine changes in expectations during the course of treatment, or
the interplay between expectations, investment, and improvement as treatment progresses. The
one exception to this is a study of self-efﬁcacy among inpatient substance abusers (Burling, et al. ,
1989), which found an association between low expectations at intake, increasing self-efﬁcacy
during treatment, and more positive outcome at the end of treatment. Clearly, this is an
important area for further inquiry.

Regarding treatment satisfaction, the ﬁndings are even more disparate. As a
characteristic presumably arising out of the client and therapist’s experience of the treatment

process, satisfaction may have important implications for subsequent treatment investment,
CXpectations, and outcome. Although the research thus far has not examined this possible
interplay, the study by Tracey (1989) provides some preliminary support. In this regard, a

curvilinear trend in satisfaction was found for both clients and therapists in cases with more

52

successful outcomes. This trend was characterized by high satisfaction at the beginning of
treatment, a drop in satisfaction at mid-treatment, and a return to high satisfaction by termination.
The author proposes that the drop in satisfaction at mid-treatment is related to a delving into more
difficult issues, a phase that is a typical mid-treatment process in successful interventions.
Statement of the Problem

Research on alcoholism has identiﬁed various risk factors occurring relatively early in
a child’s life that appear to be precursive to the development of alcohol problems and antisociality
later on. Among these, poor parenting practices, child antisocial behavior, and marital conﬂict
have all been implicated (Hawkins, et al., 1992). In addition, parental alcoholism has been
linked to youth alcohol abuse both through an association with these other risk factors, as well
as via genetic and modeling inﬂuences. A risk-focused approach, designed to extinguish or
ameliorate these precursors to alcoholism, has been advocated, with family-based, parent skills
training interventions promoted as one means of addressing some of these risk factors (Hawkins,
et al., 1992; Zucker & Noll, 1987).

Similar to the delineation of risk factors in the alcoholism literature, treatment outcome
research has begun to move away from global measures of treatment effectiveness, toward an
examination of factors associated with better or worse outcomes at termination. Regarding
cognitive-behavioral interventions in general, and parent training programs in particular, this
literature has focused largely on pretreatment client characteristics that are predictive of outcome.
While moderately informative, this approach overlooks the inﬂuence that characteristics of the
treatment process itself have on treatment success. Characteristics such as treatment expectations,
Satisfaction, and investment are likely to have important implications for treatment outcome. To
the extent that any of these treatment-related characteristics have been examined thus far, they

have typically been looked at in isolation from one another.

53

The goal of the present study is to examine the interrelationships among parent and
therapist expectations and satisfaction, parent investment, and treatment outCome in a prevention
program designed to reduce familial and child characteristics associated with the development of
alcoholism. Given the anticipation that there will be multiple pathways of inﬂuence affecting
ultimate treatment outcome, a multiple inﬂuence model is proposed (Figure 1). This conceptual
model, presented in simpliﬁed form, proposes a chain of inﬂuence among parent and therapist
expectations, parent treatment investment, parent and therapist satisfaction, and treatment outcome
as measured by changes in both parenting behavior and child behavior. Both client and therapist
expectations are anticipated to affect treatment outcome indirectly, through an association with
treatment process characteristics. In particular, it is expected that ongoing parent treatment
investment will mediate the hypothesized relationship between initial expectations and treatment
outcome. In addition, it is hypothesized that expectation and satisfaction ratings at mid—treatment
may have important implications for subsequent investment and outcome.

The model also proposes that pretreatment levels of parent psychopathology, family
economic resources, child behavior, and parenting style will be associated with treatment process
and outcome characteristics. Higher levels of parent psychopathology and greater socioeconomic
disadvantage are expected to be associated with more negative expectations about the treatment
protocol for both parents and therapists. Further, it is anticipated that level of child behavior
problems at pretest will be associated with parents’ initial expectations about the treatment
program, with more problematic child behavior associated with more negative pretreatment
CXpectations. This latter hypothesis is based on an expectation that parents will perceive the
intervention as being focused primarily on changing child behavior. In contrast, therapists may

be more likely to focus on pretreatment parenting characteristics in anticipating families’ potential

Success in the intervention. Therefore, therapists’ expectations about the potential success of the

38088 2253.: wEOan

5 85528.35 $805 3258: 98 8258505 :85»: $3.82: on. wﬁanB ABE—95$ BEE 338a 5258: 3885 I. __ 0. 5mm

 

noduoouuauom ununouona

 

 

 

unaduouuomxu unﬁnouona A tttttttt naoauauoonxﬂ uuwnouonh

;

uﬁim maﬁa-nan ........................ «33m maﬁa-unmv

 

     

 

 

 

 

 

 

 

uoa>anom OH«AU ...................... uoa>anom uuasu

 

 

 

 

 

hmoHonuuaonohnm vacuum v

m\\\ acousonom kHﬂadh

0305 OOHGH ”QOEGOHH. UQOHGQ

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

coauoouuauom vacuum <
unowuuuoonxu vacuum A IIIIIII uaoauouoonxu vacuum
AﬂmoEmomv
8:582 080330 augment moﬂmmuouoemno wwoookm “cogmouh. momumtouoagnv “coauaoboum

Vm

55

program are anticipated to be related to their initial impressions of the parents, with more
impaired parenting associated with more negative expectations on the part of the therapist.

Also of interest in the present study is the inﬂuence of parent psychopathology and family
resources on early levels of child behavior and parenting practices, as well as with pretreatment
expectations and subsequent outcome. In particular, higher levels of psychopathology and fewer
family resources are anticipated to be associated with higher levels of negative child behavior,
lower levels of positive child behavior, more impaired parenting skills, and lower expectations
for success in the program at pretreatment. Regarding outcome, changes in child and parenting
behavior are predicted to be most strongly tied to treatment process factors, with less direct
inﬂuence by these parent and family characteristics. Thus, treatment process characteristics are
anticipated to make a unique contribution to the prediction of treatment outcome, with
pretreatment parent and family characteristics controlled.

Hypotheses

Hypothesis 1. Higher pretreatment levels of parent psychopathology and fewer
pretreatment family socioeconomic resources will be associated with a) more problematic child
behavior, b) more impaired parenting, and c) more negative treatment process characteristics over
time.

Hypothesis 2. Lower parent and therapist treatment expectations at pretreatment will be
associated with a) more problematic child, parent, and family functioning at baseline, and b)
lower parent investment during Phase 1 of the intervention.

Hypothesis 3. Parent and therapist satisfaction ratings (including expectations for future

Change) at mid-treatment and termination will be positively associated with level of prior
treatment investment as well as subsequent investment (i.e. from mid-treatment to termination).

Hypothesis 4. Parent and therapist satisfaction at mid-treatment and end-of-treatment time

56

. points will also be positively associated with a) change in child behavior and b) change in
parenting style at the same time points.
Hypothesis 5. Changes in a) child behavior and b) parenting style at mid-treatment and
termination will be positively associated with prior treatment investment at the same time points.
Hypothesis 6. Treatment process characteristics (expectations, satisfaction, and
investment) will be positively predictive of improvement in child and parenting behaviors at
termination. Further, this association will be maintained when pretreatment parent, child, and

family characteristics are controlled.

CHAPTER H
METHODS
Participants

Participants were 29 intact families who completed the Michigan State University Multiple
Risk Child Outreach Program (Maguin, et al., 1994; Zucker & Noll, 1987, Zucker, et. al, 1985),
a program designed to reduce conduct problems by way of consistent monitoring and discipline,
and the reinforcement of prosocial behavior. The more long-term goal was to reduce the rate of
conduct disorder and aggression among the children, with the hope of also reducing later
delinquency, antisocial behavior, and substance abuse. In addition, the intervention was designed
to improve marital problem-solving strategies and to begin to address alcohol abuse within these
families.

The target group was comprised of young male children who are known to be at elevated
risk for the development of conduct disorder because of their membership in families with an
alcoholic father. A total of 104 families were recruited for potential inclusion in the treatment
design (including a no-treatment control group). These families were recruited primarily from
local district court records of drunk driving arrests, although a small minority of alcoholic fathers
were not identiﬁed through the court system, but were uncovered during canvassing efforts to
locate matched community control families for the larger parent project (I_I=10 in the larger
sample, p=4 for the current sample). The latter families were included in the program provided
that these fathers also had a history of DWI conviction during the lifetime of the target child.

The fathers identiﬁed for the current study were all apprehended for DWI with a blood
alcohol level (BAL) of at least .15 %. Court-identiﬁed families who gave permission to have their
names released to our project personnel were visited for an initial screening for suitability. A

BAL of .15% in itself suggests that these men had developed tolerance, but additional drinking-

57

58

related information was gathered from these fathers in order to ensure that they met the Feighner
(Feighner, et.al, 1972) diagnostic criteria for probable or deﬁnite alcoholism. The measures of
drinking behavior used for this screening will be described in more detail further on.

In addition to meeting alcohOl-speciﬁc criteria, families in this sample all had a male child
between the ages of 3-0 and 6-0, and this child and his biological mother were living in the home
at the time of recruitment. Finally, only Caucasian families were contacted, given the very low
incidence of ethnic minorities in this recruitment population, and because of the known variation
of alcohol and other problems with ethnicity. Thus, to include such heterogeneity without being
able to evaluate it was deemed less useful than to eliminate this source of variability.

Although all of the fathers met criteria for a diagnosis of alcoholism, few self-identiﬁed
as alcoholic. Thus, the father’s alcoholism was typically not made explicit at the outset of the
intervention, but commonly arose during the work with the family. It should also be noted that
52 % of the men in the current sample may be classiﬁed as antisocial, early onset alcoholics using
Zucker’s (1987) typology. Therefore, the group of alcoholic men recruited through the outreach
protocol is more representative of the subset of alcoholics who are known to be the most
damaged, with the most antisocial comorbidity, the earliest onset of alcohol dependence, and with
the greatest likelihood that substance abuse will develop among their offspring (Zucker &
Fitzgerald, 1991).

All families in the study received a fee for taking part.

Smle characteristics. A total of 104 families were initially assigned to an intervention
or a control group. The intervention protocol involved two formats: (1) a Mother-Only condition,
in which only the mother (as the primary caretaker) participated in the training program, and (2)
3 Both Parents format which involved both mothers and fathers in the intervention. These two

treatment conditions were incorporated in order to examine potential differences in child outcome

59

related to the observation that fathers are commonly less involved in managing child misbehavior
in conduct disordered families. Other research on this sample (Maguin, 1991; Zucker, et. al.,
1989) has already found the random assignment of these families to be successful, in that no
initial differences in levels of parent psychopathology, family socioeconomic status, or child
behavior problems were found among the families in the Mother Only, Both Parents, and Control
conditions. Initial comparisons of outcome between the two treatment conditions indicated that
families in which both parents participated experienced greater success in increasing their son’s
prosocial behaviors than did families in which only the mother participated (Maguin, et al.,
1994). However, subsequent analyses determined that maternal treatment investment, but not
treatment condition, differentially accounted for variation in treatment outcome (Nye, et al., in
press).

At the time of the initial screening evaluation, treatment families were randomly assigned
to one of the two intervention formats, and were then invited to participate in a program "to
improve parent-child communication patterns”. Of the 69 families initially assigned to one of the
intervention formats, 17 families become ineligible to receive or to continue with the program.
Intervention families remained eligible for continued participation provided that they remained
intact and lived within a 30 mile radius of the university ofﬁce. The "intactness" stipulation was
based on an expectation that marital dissolution might signiﬁcantly alter the homogeneity of the
sample and would potentially confound an examination of within-group trends and across-group
comparisons. This stipulation was particularly crucial for families in the Both Parents condition,
as it was deemed necessary that these parents both be actively involved in parenting the target
child. Nine of these ineligible families separated before the intervention was offered, 4 separated
during the intervention, and 4 became ineligible because they moved from the area.

Among the remaining 52 still eligible for the intervention, 7 refused to participate in the

60

program, and 16 families prematurely terminated their involvement in the intervention protocol.
Thus, the 29 families on which this study is based are all of the families who remained intact
during the time the treatment was being offered (i.e. for about one year) and who also completed
the entire intervention program (p=16 for Mother Only and p=13 for Both Parents conditions).
Intervention Procedures

The intervention program was a modiﬁcation of the Social Learning Therapy protocol
developed at the Oregon Social Learning Center to modify antisocial behavior in socially
aggressive children (Patterson, Reid, Jones, & Conger, 1975). This approach has been effective
in reducing problem behaviors among antisocial and conduct-disordered children (e. g. Kazdin,
1987; Patterson, Chamberlain, & Reid, 1982). Further, this type of intervention has been
identiﬁed as potentially useful in preventing substance abuse by targeting child antisocial behavior
and poor parenting practices, both of which are associated with later problems with alcohol
(Hawkins, et al., 1992; Zucker & Noll, 1987). This intervention program differs from other
parent training approaches (e. g. Dumas & Wahler, 1983; Webster-Stratton, 1984) in regard to
its scope and underlying orientation. Because marital conﬂict has been identiﬁed as a precursor
to substance abuse problems (e.g. Baumrind, 1983; Penning & Barnes, 1982), a disrupter of
parenting (e. g. Gottman & Katz, 1989), and an impediment to success in parent training
programs (e. g. Dadds, Schwartz, & Sanders, 1987), the protocol used here incorporated a marital
problem-solving component in the treatment regimen, rather than adhering to a strict child-
focused paradigm. Family conﬂict and marital dissatisfaction were frequently encountered during
sessions conducted with our sample, and often arose in relation to the fathers’ excessive drinking.
As a result, speciﬁc attention was routinely paid to parents’ alcohol and drug problems and
marital functioning as part of the intervention.

This program was explicitly designed as an alcoholism prevention program for children

61

in high-risk circumstances. Therefore, in contrast with OSLC interventions, target families in
this study were recruited not because of identiﬁed child behavior problems, but because of the
multiple risk factors associated with growing up with an alcoholic parent. Thus, the present
version was identiﬁed more as education than therapy, and was offered through outreach efforts
targeting a high-risk community population of families in which the target child had typically not
yet come to public attention for his disruptive behavior. The children in our study were also
signiﬁcantly younger (mean age of 4.4 years) at the beginning of treatment than those children
involved in OSLC’s programs. Because these families were recruited via an outreach protocol
targeting a designated high-risk population, the program was not presented as an offer of
treatment, but as an educational program designed to enhance parent-child communication and
improve parent-child relationships. Therefore, the work necessarily focused on ways to enhance
parenting techniques and child-rearing practices in order to prevent later child behavior problems.
Further, because parents were neither self-selected nor court-referred, some needed to be
convinced of the usefulness of the program.

The intervention was designed to be a ten-month-long interaction of approximately 28
sessions with each family, and proceeded in two phases. Phase I focused on the introduction and
application of child management skills through weekly sessions with families and two between-
session telephone contacts. Phase I ended when the family had successfully acquired these skills,
and/or when 12 to 16 sessions had been completed (approximately four months into the protocol).
During Phase 2, biweekly sessions and weekly telephone contacts were conducted to support and
reinforce the child management techniques while beginning to more speciﬁcally address
signiﬁcant marital and other family issues. Although the physical setting identiﬁed for the
intervention was the Clinical Center at the university, about 45% of the families received some

or all of the intervention program in their homes. The treatment staff was comprised primarily

62

of post-masters level doctoral graduate students in clinical psychology, as well as two masters
level social workers. All therapists received 20 hours of training related to the treatment protocol
before beginning the work, and four hours of weekly group supervision with one of two fully
licensed clinical psychologists (RAZ and RBN). In addition to monitoring the adequacy of
program implementation, supervision routinely focused on issues of therapist countertransference
and other treatment process issues, as these might impede successful application of the regimen.
Within the current sample of families, eleven therapists are represented.

Although the intricate details of this program are best described elsewhere (Zucker, Noll,
Cruise, Kriegler, Wehner & Mitchell, 1985) a brief overview of the procedures is provided here.

Baseline data collection and initi_al interview. The program was initiated with the
collection of baseline data on the target child’s behavior. The initial intervention interview was
then conducted with the parent(s) and the target child, and was designed to 1) establish a
relationship between the therapist and the family, 2) engage the parent(s) in observations of the
child’s behavior (pinpointing and tracking), and 3) schedule times for the therapist to contact the
parent(s) during subsequent weeks. Participating parents were given the assignment of observing
the target child’s behavior for one hour per day. This tracking was done by each parent who was
involved in the intervention program, and "minding" and "not minding" behaviors were
pinpointed as the focus of these observations. These behaviors were tracked, labeled for the
child, and recorded by the parent(s). During the subsequent week, the therapist contacted the
parent(s) by phone to discuss the incidence of "minds/not minds" and to address whatever issues
needed to be addressed.

Tracking and initial point contract sessions. In the next session(s), the therapist further
taught the parent(s) how to attend to/observe the child’s behaviors. For each tracking segment,

a particular problem behavior and its prosocial opposite were observed and counted. Phone

63

contacts were maintained between the therapist and the parent(s). In subsequent sessions, the
therapist introduced and described point contracts to the parent(s). The contract is central to
social learning treatment, and was used to teach family members to negotiate agreements among
themselves. The ﬁrst contract was the point contract or star chart, and focused on the positive
aspects of the target child’s behavior. This contract included one or two chores for the child to
do and one or two child behaviors that the parent(s) would like to see increased. Once prosocial
behaviors and chores were selected and identiﬁed, point values were assigned to each and a menu
of potential rewards was generated, with a criterion set for obtaining each reward. At the end
of each day, the parent(s) reviewed with the child the points that he earned and/or forfeited .
Daily and weekly rewards were delivered, with appropriate social reinforcement. With the star
chart, the focus was on "catching the child being good", and stars were administered as token
reinforcement for positive behaviors. Rewards were given after a speciﬁed number of stars had
been earned by the child. Again, the therapist called the families between sessions to determine
progress and identify and deal with problems.

Time Out sessions and the familiarization home visit. Once the star chart or point
contract had been successfully implemented in the home, the therapist discussed discipline
practices and philosophies with the parent(s), including what forms of discipline the families had
been most comfortable and satisﬁed with. Then the "Time Out" procedure was introduced, and
the parent(s) viewed the "Time Out!" ﬁlm (Northwest Family and School Consultants, 1981)
before continuing with this topic in the next session. In session "B", the therapist reviewed and
ran through the Time Out procedures, discussing deﬁnition, implementation, and the speciﬁc
steps necessary to carry the procedure out. After thorough discussion and explanation, the
therapist typically asked the parent(s) to practice Time Out with the child and to clearly explain

the process to him before the procedure was actually put to use. The therapist then scheduled

64

a time to visit the entire family at home in order to examine the designated Time Out space, to
get a sense of the physical environment in which the family operates, and to provide the family
with the opportunity to welcome the therapist into the home. Phone contacts between the
therapist and parent(s) were also maintained.

Problem solving. The next stage involved the introduction and practice of communication
skills involved in effective problem solving. The three basic components - active listening,
generating new alternatives, and evaluating techniques - were taught over the course of several
sessions. As with the other aspects of the program, presentation of material was accompanied
by practice within the session and by homework assignments to be carried out between sessions.
Although the speciﬁc procedures during the problem solving session varied somewhat depending
upon whether both parents were present or only the mother was involved, the basic components
and strategies were the same in both conditions. The goal was to introduce problem solving skills
that would assist the family in maintaining the skills that had already been taught in earlier
intervention sessions, by improving the communication and conﬂict-resolution abilities of the
family members. Given the young age of the children, the problems most commonly dealt with
in this segment of the program were marital rather than being more broadly family-focused. As
always, between-session phone contacts were made so that the therapist could discuss progress
and problems with the parent(s).

Other help needs and end phase (termination) work. Because families with an alcoholic
parent tend to have multiple problems and difﬁculties, the parent(s) often brought up family
problems that did not directly involve the target child. The therapist was prepared to attend to
and acknowledge these difﬁculties, and to engage in interventions that would be helpful to the
family. At the same time, the therapist worked to keep the desired continuity of the intervention

program’s regimen. However, the pace at which a family progressed through the program may

65

have been slowed because of other family difﬁculties. Further, not all family problems were
expected to be solved at the end of the planned 28 week intervention. Therefore, additional
referrals for further assistance may have been necessary at the end of the intervention, in order
to help these families deal with life issues other than those involving the child(ren).
Data Collection Procedures

Each family involved in the study completed a standard battery of questionnaires and
participated in a variety of interviews and direct observation sessions (Zucker, Noll, &
Fitzgerald, 1986; Zucker, et. al., 1985). Data were collected during the course of an eight
session contact schedule, as well as during the intervention program. The data used in the
present study were collected as close to the onset of the intervention phase as possible, and at
mid-treatment and end-of-treatment time points. Additional follow-up data were collected from
the parents six months subsequent to the termination of treatment. The ﬁrst set of post-test
instruments was collected in the ﬁfth month of the program, after the families had begun meeting
with the therapist on a bi-weekly basis. These data were collected in the family’s home, by a
trained team comprised primarily of graduate students. Measures related to aspects of the
intervention work and to observed child and parent functioning were completed by the therapist
assigned to each family. These therapist ratings were completed immediately after each
intervention session or other family contact, as well as at the scheduled post-test intervals.
Measures

The measures that are of relevance for this research assessed parents’ perceptions of their
child’s behavior problems, as well as levels of drinking, antisocial behavior, and depression in
the parents. Measures of family sociodemographic characteristics were also included.
Information about parents’ expectations about the treatment program and their satisfaction with

various aspects of the intervention was collected, and therapists were also asked to rate their own

66

expectations of the potential success of the program, and their satisfaction with how the family
was progressing during treatment. In addition, therapist ratings of child behavior, parenting
style, and parent cooperation/investment throughout the intervention were collected. Finally,
independent observations of parenting behavior were conducted prior to treatment, as a more
objective measure of initial parent functioning against which to validate therapists’ ratings.
Pretreatment Measures: Parent Psychopathology

Drinking measures. Several instruments assessing alcohol and drug involvement were
administered individually to the parents in the study. Information provided by these measures
includes current alcohol use, alcohol-related problems, and drinking history. Parents completed
a detailed Drinking and Drug History (Zucker & Noll, 1980b), the Short Form of the Michigan
Alcoholism Screening Test ([SMAST]; Selzer, 1975), and were interviewed about their drinking
practices as part of the Diagnostic Interview Schedule ([DIS]; Robins, Helzer, Croughan, &
Ratcliff, 1980). The present work focused on information regarding the variety and duration of
reported drinking problems, and the age at which the respondent reports having gotten drunk for
the ﬁrst time. These data were used to compute the Lifetime Alcohol Problems Score ([LAPS];
Zucker, 1991), an index of the extent of alcoholic involvement and problems over the individual’s
life course. The component subscores have been standardized within our project sample, and
validity studies have shown LAPS to be an adequate index of the extent of alcohol-related
impairment (Zucker, 1991). In the present study, this score was chosen as an indicator of
parents’ lifetime alcohol-related trouble prior to treatment. A measure of current alcohol
consumption was go; included in the present work because of evidence (e. g. Nye, 1992; Reider,
1991; Reider, et. al., 1989) that current alcohol use is of limited predictive validity among these
families, primarily because of a different pattern of relationships with LAPS and antisociality

(ASB) for fathers than for mothers (r= -.27 and r= -.03 respectively for fathers, and r= .41 and

67

r= .40 for mothers). A likely explanation for this discrepancy is that all of the men in the study
were arrested for DUI prior to initial data collection. Thereafter, many of them quit drinking,
thus making reported current alcohol consumption an inaccurate indicator of trouble with alcohol.

Depression. Current parental depression was measured with the Short Form of the Beck
Depression Inventory ([BDI]; Beck & Beck, 1972), which contains 13 groups of statements
addressing various areas of functioning that are often affected by depression. Scores on this short
form correlate between .89 and .97 with the long form, and the psychometric durability of this
measure is well-documented (Beck, Steer, & Garbin, 1988). Level of Worst Ever Depression
(Zucker & Fitzgerald, 1992) was generated from the Diagnostic Interview Schedule, via the
Hamilton Rating Scale for Depression (Hamilton, 1960). The interviewer uses this scale to
generate an index of the time in the individual’s life when s/he was most depressed. Both the
Beck and Hamilton Worst Ever Depression scores were used as other indices of trouble among
the parents in this sample.

Antisocial behavior. Parental antisocial behavior was measured with the Antisocial
Behavior Checklist (Zucker & Noll, 1980a). This instrument explores the occurrence of
antisocial behaviors in the respondent’s childhood, adolescence, and adulthood, in relation to nine
different content areas (e.g. trouble with the law, job-related antisocial behavior). Parents were
asked to indicate the frequency of their participation in various antisocial behaviors.
Psychometric properties of the instrument have been found to be sufﬁcient, as test-retest
reliability is strong (.91 over four weeks) and internal reliability is high (coefﬁcient a= .93)
(Zucker & Noll, 1980c).

Parent psychopathologv index. In the present work, lifetime and current levels of parent
symptomatology were not analyzed separately due, in part, to the limited utility of using the

current alcohol consumption measure with this particular sample of families, as well as the

68

observation of high intercorrelations among the lifetime measures and the BECK index of current
depression. In fact, all of the parent psychopathology indices have been found to be sufﬁciently
intercorrelated as to warrant combining them into general mother and father psychopathology
scores (e. g. Nye, et al. , in press). Therefore, the present study used a composite score for parent
trouble made up of LAPS, ASB, HAMW, and BECK (see Appendix A). This is consistent with
other research from the parent project (Cruise, 1991; Reider, 1991) which identiﬁed a composite
measure of parent psychopathology to be a better summary index of parent functioning than were
each of the individual measures.
Pretreatment Characteristics: Family Resources

Sociodemographic variables. Parent and family sociodemographic variables related to
parent education and income, and family socioeconomic status were also included in analyses.
Earlier research with this group of families (Nye, 1992) identiﬁed a composite measure of family
resources that was inﬂuential in predicting eventual child outcome. This Family Resources
cluster, comprised of mother and father education and income, and family socioeconomic status
(as measured by the Duncan TSEI2 index (Stevens & Feathennan, 1980)), will be used in the
current work, as well. Family SES was computed as the average of both mothers and fathers’
occupational status when both parents were employed, and on the occupation of the employed
parent when only one parent was working outside of the home. When neither parent was
employed, family SES was assigned a standard score below the lowest occupational code on the
scale. Correlations among the variables in the Family Resources clusters are listed in Appendix
B. Parent psychopathology and family demographic characteristics of the treatment sample are
presented in Appendix C.

Pretreatment Characteristics: Child Behavior and Parenting
Child behavior. The Child Behavior Rating Scale - Preschool Version (CBRS) (Noll &

69
Zucker, 1985; adapted from Hopps, 1985) is designed to measure both prosocial behaviors (49

items) and undesirable, coercive behaviors (35 items) typical of preschool children. Items are
rated on a 7-point Likert-type scale, ranging from "Never" to "Always". Mothers and fathers
completed independent ratings "of child behavior during one of the pretreatment data collection
visits. The therapist also completed the rating scale, based on his/her observations of the target
child’s behavior during the initial contact(s) with the family. From these ratings are generated
two scores reﬂecting current child behavior: Positive Behavior is the average frequency of all
prosocial behaviors reported, and Negative Behavior is the mean rate of undesirable/antisocial
behaviors. Adequate test-retest reliability has been established, and the measures correlate
substantially with external validity indicators (Nye, 1992; Maguin, et al., 1994).

The CBRS was also completed by both parents at the mid-treatment (Post-test 1),
termination (Post-test 2), and follow-up (Post-test 3) data collection visits. In addition, therapist
reports on child behavior are available for the mid-treatment and termination time points. This
longitudinal data on child behavior will be used as one of the treatment outcome measures (see
below).

Parenting. The Observer Impressions Inventory, used extensively by the OSLC group,
is a measure for reporting various aspects of Observed interactions between parents and the target
child. These observations were made and rated by the therapist after her/his ﬁrst contact with
the family. Included in this measure are 5-point ratings of discrete observed parent-child
interactions, as well as more global ratings of parents’ disciplinary style (ranging from "doesn’t
ﬁt the parent at all" to "ﬁts the parent perfectly) and general manner of interacting with the target
child. These more global ratings include the extent to which mothers and fathers appear to act
in overly strict, overly permissive, erratic, and/or consistent ways with the target child. Global,

rather than discrete, parenting behaviors were used in analyses because of evidence that global

70

parenting style moderates the inﬂuence of speciﬁc parenting practices on child functioning
(Darling & Steinberg, 1993), and because of concern that ratings of discrete behaviors might be
more situation or context-speciﬁc. A principal components analysis of the global parenting items
identiﬁed four factors characterizing parenting style in relation to the target child. The primary
factor identiﬁed consisted of 14 items and accounted for 40% of the variance. The other factors
accounted for an additional 26% of the variance. The items loading on this factor were
commensurate with what might be characterized as non-coercive or authoritative parenting, and
included such positive indicators as "Is a positive and reinforcing parent", "Treats target child
with respect", and "Is consistent, even-handed, ﬁrm when necessary". Items loading in a
negative direction on this factor included "Is overly permissive, laissez-faire, negligent", "Uses
nagging or nattering to get compliance", and "Expresses anger/hostility while disciplining". The
14 items on this factor were combined into an averaged composite score representing level of
"Authoritative Parenting". Reliability for this scale was determined to be adequate (coefﬁcient
a== .90). These items, and their factor loadings, are presented in Appendix D.

Because therapists were the only raters of parenting style over time, the validity of
therapist reports was assessed by comparing items in the Authoritative Parenting score with other
observational ratings of parenting behavior at baseline. Videotaped dyadic play sessions between
each parent and the target child were conducted prior to the beginning of treatment, and involved
each parent and the target child in three 10 minute situations requiring increasing amounts of
parental control (Zucker & Fitzgerald, 1991). The ﬁrst segment was child-directed play, during
which the instructions were for the parent to let the target child select the play activity. Parent-
directed play, where the parent selected the activity, comprised the second 10-minute segment,
followed by a clean-up interval during which the parent was expected to ask the child to put away

the play materials (Whipple, et al., 1995). A revised version of the Belsky Coding System for

71
Parent-Child Interaction (Belsky, Youngblade, Rovine, & Volling, 1991) was used to characterize

the nature of these parent-child interactions. Coding was conducted by seven graduate and
undergraduate students, who received a minimum of 45 hours of training, maintained 75 %
reliability, and were blind to other information about the families (including the fact that they
were part of the intervention program).

The revised coding protocol (Whipple, Denburg, & Davies, 1992) includes 9 parent
scales assessing level of warmth or affection, and type of control. These were correlated with
items from the Authoritative Parenting scale of the Observer Impressions Inventory. Signiﬁcant
correlations are presented in Appendix E. The Authoritative Parenting factor was examined as
a marker of pretreatment parenting style in the current study, and as an index of treatment-related
change based on therapist observations made at the mid-treatment and termination time points (see
below).

Treatment Process Measures

Treatment expectations. Prior to the commencement of treatment, parents and therapists
were asked to rate their expectations about the intervention program. Speciﬁcally, parents were
asked how likely they felt the program would bring about change in the target child, ranging from
very likely to very unlikely (on a 5-point scale). Similarly, therapists reported their own
expectations that the program would bring about change in the target child, based on their
knowledge of the family. During post-test data collection visits, parents were asked to indicate
their expectations that the program would produce further change or maintain already produced
change in the target child, using the same 5-point scale. Therapists also reported their
expectations for further change in child behavior, at the mid-treatment and termination intervals.
These post-test expectation ratings were found to correlate signiﬁcantly with satisfaction ratings

provided by parents and therapists at the same time points. Therefore, parent and therapist

72
expectations were included in the composite Satisfaction scores for post-test data (see below).

Treatment investment. Treatment Investment is operationalized as the amount of parent
cooperation with and receptiveness to the treatment regimen exhibited throughout the intervention
process, as reﬂected by both within-session resistance and between-session noncompliance with
homework assignments. In session-by-session analyses, Patterson and Chamberlain (1988) have
found that these two indices of treatment cooperation or investment are positively related to the
extent to which change in child-rearing practices and child behavior takes place. They also
suggest that these factors are crucial to the understanding of individual differences among families
in their responsiveness to treatment. Further, these dimensions have been found to covary with
family disrupters such as parent psychopathology (e. g. Nye, 1992).

For the purposes of the present study, parent investment was measured by three
instruments developed by the Oregon Social Learning Center group: Weekly Homework Ratings,
the Therapist-Client Cohesion Scale, and the Client Involvement Rating. The Weekly Homework
Ratings were designed to indicate the extent to which the parent(s) attempted to carry out the
homework assignments, and ratings are based on a 5-point scale, ranging from "tried very hard"
to "didn’t try at all". The Therapist-Client Cohesion Scale is an index of the therapist’s personal
reactions to the parent(s) during each session. The therapist indicated how s/he felt about the
parent, based on a 5-point range from "felt very positively toward" to "was very irritated with".
The therapist further indicated his/her sense of how well s/he worked with each parent, ranging
from "very well" to "poorly". The Client Involvement Rating is also a series of 5-point scales
on which the therapist rated each parent on the amount of verbal involvement, interpersonal
withdrawal, openness to new ideas, hostility or friendliness, and overall resistance or helpfulness
exhibited during the sessions. All three instruments were ﬁlled out by the therapist after each

session with the family. Reliability for these measures has been found to be adequate (Nye,

73
1992).

Other analyses on this data set identiﬁed substantial colinearity among the parent
involvement, cohesion, and homework items, indicating that the use of a single investment cluster
based on these three dimensions is appropriate (Nye, 1992). Therefore, cluster scores based upon
a cumulation of data from Pretest through Post—test 1 (Early Investment), one cumulating
treatment investment from mid-treatment through termination (Late Investment), and one
summarizing investment across all sessions (Cumulative Investment) were used in the present
analyses. Because fathers were not always included in the intervention program, the "Parent
Investment" score was based on mothers’ treatment investment scores in the Mother-only
condition, and on the average of mother and father scores in the Both Parents group.

Treatment satisfaction. Treatment satisfaction measures were completed by parents and
therapists at mid-treatment and upon completion of the program. These instruments present each
respondent with a 5-point scale on which to rate: changes in child behavior (from much worse
to much better); feelings toward the target child ("much more negative" to "much more
positive"); the effect of the work on the marital relationship (from "very negative" to "very
positive“), and the effect of the work on both the parent-child relationship and on each parent
personally (ranging, again, from "very negative" to "very positive" effect). In addition, parents
were asked to indicate how well the therapist seemed to understand problems or difﬁculties
brought up in sessions (ranging from "doesn’t understand at all" to "always understands"), and
how well they liked the therapist ("not at all" to "very much"). Similarly, the therapist form asks
for ratings of the parents’ ability to understand treatment directions and techniques brought up
in sessions, and how well the therapist liked each parent individually. Finally, the measure also
includes a 7-point global rating of parent and therapist satisfaction with the treatment program,

ranging from "very dissatisﬁed" to "very satisﬁed".

74

In order to reduce the number of variables being analyzed, a single mean Satisfaction
score was generated across all items on each satisfaction inventory. As with the treatment
investment mean, Parent Satisfaction scores were based on mothers’ mean satisfaction ratings in
the Mother-only condition, and on the average of mother and father ratings in the Both Parents
group. These Satisfaction means also included parent and therapist post-test expectations about
the future efﬁcacy of the program, because of substantial correlations with the satisfaction items
(Appendix F). Reliability for these satisfaction scales was found to be adequate, with
standardized coefﬁcient alphas of .81 for parent satisfaction ratings, and .89 for therapist ratings.
Treatment Outcome Measures

Child behavior problems. As described above, the Child Behavior Rating Scale -
Preschool Version (CBRS) (Noll & Zucker, 1985; adapted from Hopps, 1985) measures the
extent to which the target child engages in both prosocial and undesirable or coercive behaviors,
ranging from "Never" to "Always". This scale has already been used successfully to identify
child behavior change in the M.S.U. intervention program (Maguin, 1991; Nye, et al, in press;
Zucker, et. al. , 1990). In the present study, child outcome was assessed via changes in the rates
of both prosocial and negative child behavior from pretreatment to post-test levels. Correlations
among mother, father, and therapist ratings of the target child’s behavior are presented in
Appendix G. An examination of these indicates that no two observers rated the target child’s
behavior consistently with one another. This is in line with other comparisons of mother, father,
and therapist ratings of child behavior, which suggest that idiosyncratic variations typically occur
for each rater (Patterson, Dishion, & Chamberlain, 1993). Therefore, the best estimate for child
behavior was considered to be the average of all ratings, thereby reducing the contribution Of
idiosyncratic variance.

In the present study, mother, father, and therapist reports on the Child Behavior Rating

75

Scale were combined to generate an averaged index of child behavior at the pretreatment, mid-
treatment, and termination time points. Because therapists were not able to make ratings of the
target child’s behavior at the six—month follow-up, only the mean of mother and father ratings
were used for that post-test period. Correlations between the averaged child behavior scores and
the individual mother, father, and therapist ratings were generally found to be substantial for all
independent ratings, and incremental over the individual contributions of each rater. Thus,
averaging child behavior means across raters was found to yield a meaningful summary score that
represents reports from mother, father, and therapist (Appendix F). Further, analyses based on
the larger treatment sample found this summary score to be correlated with coders’ impressions
of the target child during the videotaped parent-child interaction task. In this regard, coders’
rating of how much they liked the child in the videotape was associated with both Positive and
Negative Child Behavior means at baseline (52: .22, p< .05 for positive behavior, and g2=-.25,
p < .01 for negative behavior). The pretreatment positive child behavior mean was also correlated
with therapists’ ratings of the target child’s conduct on the Observer Impressions Inventory
([2: .44, p< .01). Correlations between the composite Positive and Negative Child Behavior
means were: g2=-.07, p=n.s. at baseline, r2=-.51, p_< .01 at Post-test 1 (mid-treatment), £2:-
.56, p_< .01 at Post-test 2 (termination), and f=-.71, p_< .001 at Post-test 3 (follow-up).
Parenting style. As described earlier, the Observer Impressions Inventory was used to
generate an index of Authoritative Parenting style, based on therapists’ observations of
interactions between parents and the target child. These observations were made and recorded
by the therapist at the ﬁrst session, and at the mid-treatment and termination time points.
Changes in therapists’ ratings of each parent’s disciplinary style represent the second outcome

measure of interest in this study.

CHAPTER HI
RESULTS
Data Analytic Procedures

Prior to analysis, each set of data was examined for outliers. Outliers were deﬁned as
extreme values falling outside of a normal curve within a given variable’s frequency distribution
histogram. These were then corrected by assigning the value adjacent to the closest non-outlying
variable, thus preserving the rank ordering of variables. Missing data were estimated using
individual baseline parent and family variables which were signiﬁcantly correlated with the
variable to be estimated. Overall, this technique was used for less than 8% of the data set. The
only exception to this was for parent and therapist pretreatment expectations, a signiﬁcant portion
of which (30-50%) were missing because the instruments measuring these were not added into
the initial assessment battery until after some of the families had already begun treatment.
Nonetheless, missing data points for these instruments were also estimated from parent and family
baseline characteristics because of an interest in examining the relationship between pretreatment
expectations and subsequent process aspects of treatment. Conclusions drawn from analyses
involving pretreatment expectations must therefore be considered to be quite tentative due to the
amount of missing data that were imputed.

Although typical of treatment-based studies, the small sample size raised additional data
analytic considerations. Both the complexity of the conceptual model (see Figure 1) and the large
number of variables relative to the number of participants precluded a path analysis of the full
treatment process model. Instead, separate analyses were performed on selected smaller
components of the model in order to examine the relationships of interest. This necessarily
resulted in a greater overall number of analyses than would have been conducted had the sample

size been able to accommodate an examination of the entire model. Although the number of

76

77

analyses extends beyond optimal parameters given the size of the sample, the majority of the
preliminary analyses were conducted in order to maximize the possibility of subsequently
rejecting the central hypotheses. In this regard, relationships involving pretreatment parent,
family, and child characteristics - which were hypothesized to be more distal and indirect
inﬂuences on treatment outcome -- were thoroughly examined during preliminary analyses in
order to ensure that relevant pretreatment predictors of outcome might be adequately controlled
prior to testing the contribution of treatment process characteristics. In order to be able to more
fully explore and adequately identify these relationships, one-tailed rather than two-tailed tests
were used to evaluate some of the initial hypotheses regarding the relationships among
pretreatment parent," child, and family characteristics, pretreatment parent and therapist
expectations, and treatment process characteristics, and in examining the relationship between the
child and parent treatment outcome measures.

For all analyses testing the proposed hypotheses, an alpha level of .05 was considered
signiﬁcant. Additional exploratory analyses examined potential interactions between pretreatment
parent and child functioning and treatment process characteristics in predicting treatment outcome,
as such relationships had been previously identiﬁed in research with the larger treatment sample
(Nye, 1992). For these latter analyses, a more lenient 10% alpha level was used because of the
restricted sample size and the exploratory nature of these tests of moderation. In general, the
results of these analyses were consistent with one another and with the previous work. However,
the less stringent alpha level and the observation that hypothesized interaction effects were not
always signiﬁcant, raised concern that identiﬁed moderator effects might be spurious. Therefore,
these results were omitted from the already complex main body of the ﬁndings, and are reported
in Appendix H, instead.

All variables were analyzed as continuous, except for the Treatment Condition construct

78
(coded as 1 for the Mother-Only condition and 2 for the Both Parents condition).

Changes in Child Behavior and Parenting Over Time

In order to establish the nature and direction of changes in the child and parent outcome
measures, preliminary analyses examined variations in the Positive Child Behavior, Negative
Child Behavior, and Authoritative Parenting means over time. Repeated measures MANOVAS
and post-hoe (Duncan) comparisons identiﬁed signiﬁcant change in child behavior between
pretreatment and post-tests. An overall group effect for Positive Child Behavior was observed
(E(3,26)= 18.54, p< .001), with positive behavior increasing from pretreatment to the mid-
treatment, termination, and follow-up time points. Similarly, this treatment group showed an
overall decrease in Negative Child Behavior from pretreatment (£(3,26) = 11.06, p_< .001) to mid—
treatment, termination, and follow-up. However, neither child behavior measure varied
signiﬁcantly between post-tests, indicating that improvement in child behavior occurred primarily
during Phase 1 of treatment and was then maintained through termination and follow-up. This
is consistent with an earlier study involving this treatment sample (Nye, et al., in press),
suggesting that the primary treatment effect for child behavior occurred during the ﬁrst, more
explicitly child-focused phase of the intervention protocol, with these changes subsequently
maintained but not increased over time. Comparisons between Authoritative Parenting scores
revealed a signiﬁcant time effect (E(2,27)= 11.52, p < .001), with parenting style becoming more
authoritative from pretreatment to termination, based on Duncan post-hoe comparisons (Table 1).
Thus, the intervention protocol effected signiﬁcant change in both child behavior and parenting
by the end of the program.
Relationship Between Child Behavior and Parenting Over Time

The relationship between the child behavior means and the Authoritative Parenting

construct was examined next, to establish whether changes in the targeted parenting practices

79
Table 1

Post hpc Comparisons (Duncam of Changes in Positive pnd Negative Child Behavior and

Authoritative Parenting

Assessment Phase

 

Pretest Post-test 1 Post-test 2 Post-test 3

Outcome Measure M Si), M SD M _S_D _M _S_D E

 

Positive Behavior 3.97a .39 4.40b .40 4.49b .41 4.60" .43 13.11***
Negative Behavior 3.02‘ .39 2.80b .39 2.63b .39 2.66b .46 5.56M

Parenting 3.35‘I .46 3.58”“b .56 3.80” .58 -- -- 5.13**

 

Note. Means having different superscripts are signiﬁcantly different at p< .05 (Duncan
comparisons).

** p< .01. *** p< .001.

80

were associated with concomitant change in the child behaviors which these parenting techniques
were expected to modify. A series of partial correlational analyses (one-tailed tests of
signiﬁcance) were conducted to test the hypothesis that level of Authoritative Parenting at mid-
treatment and termination would be associated with the Positive and Negative Child Behavior
means at the same time points, with pretreatment child behavior controlled. The relationship
between pretreatment parenting and post-test child behavior was also analyzed, with the
expectation that Authoritative Parenting at baseline would be minimally associated with
subsequent child behavior because of the changes in parenting style effected during the
intervention. These analyses revealed that level of Authoritative Parenting at mid-treatment was
correlated with changes in both Positive and Negative Child Behavior at the same post-test
(13:52, p< .01 for Positive Behavior, and f=-.60, p< .001 for Negative Behavior). Regarding
these outcome measures at termination, Authoritative Parenting was associated with Positive Child
Behavior (:2: .47, p< .01) and Negative Child Behavior (p2=-.63, p< .001) at the end of the
intervention, with more authoritative parenting associated with greater positive and less negative
child behavior. In addition, Authoritative Parenting style at termination was correlated with child
behavioral outcome at follow-up for Negative Child Behavior (_r_2=-.39, p< .05), suggesting that
increases in parents’ authoritativeness observed during the intervention also contributed to reduced
negative child behavior six months later. Pretreatment level of authoritative parenting was not
signiﬁcantly associated with either of the child behavior means at any post—test time point.
Inﬂuence of Pretreatment Characteristics

Pretreatment Characteristics and Treatment Outcome Measures

The analyses conducted thus far indicate that the intervention protocol was successful in
increasing authoritative parenting from pretreatment to termination, as well as increasing positive

child behavior and decreasing negative child behavior from pretreatment to the mid-treatment,

81

termination, and follow-up time points. Further, these changes in child behavior were associated
with observed increases in authoritative parenting during the intervention. The next set of
analyses was conducted in order to identify which parent and family characteristics at
pretreatment were associated with the observed changes in child and parent outcome. These
analyses tested the hypothesis that higher levels of parent psychopathology and fewer family
resources prior to treatment would be associated with more impaired parenting and more
problematic child behavior at all time points. One-tailed tests of partial correlations were
conducted, controlling for baseline levels of parenting and child behavior. Treatment Condition
(Mother-Only and Both Parents) was also included in these preliminary analyses, in order to
identify potential differences in outcome related to treatment participation by one versus two
parents.

Child behavior. Father psychopathology was signiﬁcantly associated with both positive
and negative child behavior at all time points, and in the anticipated directions. Greater Father
Psychopathology at pre-treatrnent was associated with lower rates of Positive Child Behavior and
more Negative Child Behavior at mid-treatment, termination, and follow-up, with pretreatment
levels of these child behavior indices controlled. Mother Psychopathology was also positively
correlated with Negative Child Behavior at all time points, but was signiﬁcantly (negatively)
correlated with Positive Child Behavior at follow-up only. Finally, the Family Resources cluster
was positively correlated with Positive Child Behavior at mid-treatment, with pretreatment
Positive Child Behavior controlled (Table 2). Treatment Condition was not signiﬁcantly
correlated with Positive or Negative Child Behavior at any of the time points.

Authoritative parenting. In examining the partial correlation coefﬁcients between
pretreatment parent and family characteristics and subsequent changes in parenting behavior, only

Father Psychopathology was signiﬁcantly correlated with Authoritative Parenting at mid-treatment

82

(_rz=-.46, p< .01) and termination (r2=-.42, p< .05), with pre-treatment parenting controlled.
Thus, greater paternal psychopathology at baseline was associated with more impaired parenting
throughout the intervention protocol. Treatment Condition was also not found to be associated
with this index of treatment outcome.
Pretreatment Characteristics and Initial Treatment Expectations

Correlations were also generated in order to examine the inﬂuence of pretreatment parent,
child, and family characteristics on pretreatment parent and therapist expectations. It was
anticipated that greater parent psychopathology, fewer family resources, and more problematic
child behavior prior to treatment would be associated with lower initial parent and therapist
expectations that the intervention would be successful. Signiﬁcance tests were again one-tailed
because the expected direction of these correlations was speciﬁed a priori. In these analyses,
pretreatment Parent Expectations were found to be negatively correlated with Mother
Psychopathology (g2=-.33, p< .05), indicating that parents in families with mothers who were
more psychologically troubled expected the intervention to be less beneﬁcial. The association
between Family Resources and Parent Expectations also approached signiﬁcance (g2= -.30,
p= .06). Pretreatment Therapist Expectations were positively correlated with Family Resources
(:2: .41, p< .05) at baseline, indicating that therapists expected the intervention protocol to be
more useful to families with more socioeconomic resources. Interestingly, neither initial child
behavior nor pretreatment level of authoritative parenting were associated with therapist
expectations, suggesting that baseline levels of those behaviors to be targeted by the intervention
did not inﬂuence the extent to which therapists thought a family would beneﬁt from the protocol.
In contrast, the correlation between parent expectations and positive child behavior at baseline
approached signiﬁcance (r2=.31, p=.051), suggesting that parents whose children initially

exhibited more positive behaviors expected to have greater success in the program. Parent and

83
Table 2
Partial Correlations Between Pretest (Pretreatment) Parent and Family Characteristics and Post-

test (PT) Positive apd Negative Child Behavior, Controlling for Baseline Positive and Negative
Child Behavior Means

 

Post-test Child Behavior Mean

 

 

Pretest Characteristic Positive Behavior Negative Behavior
PT 1 PT 2 PT 3 PT 1 PT 2 PT 3
Mother Psychopathology -. 19 -.30 -.31* .31* 35* .31*
Father Psychopathology -.40* -.46** -.38* .39* .39* .31*
Family Resources .42** .25 .26 -.05 -. 10 -.30

 

*p< .05, one-tailed. **p< .01, one-tailed.

84

Therapist Expectations at baseline were poorly related Q2: .04, p=n.s.).
Pretreatment Characteristics and Treatment Process Measures

The next set of analyses examined the extent to which baseline parent and family
characteristics were also signiﬁcantly associated with parent treatment investment, and parent and
therapist satisfaction ratings (including expectations about continued treatment progress)
throughout the intervention. One-tailed tests of Pearson correlations were conducted to test the
hypothesis that greater parent psychopathology and fewer family resources would be associated
with less parent investment, and with less parent and therapist satisfaction with treatment over
time.

Parent treatment investment. No parent or family baseline characteristics were
signiﬁcantly associated with level of Parent Investment during the ﬁrst half of the intervention
(Early Investment), although correlations with Mother Psychopathology and Father
Psychopathology approached signiﬁcance (;2=-.28, p= .07 for both). Baseline level of Positive
Child Behavior was positively associated with Early Parent Investment ([2: .47, p< .01), and the
correlation between pretreatment Authoritative Parenting and Early Investment also approached
signiﬁcance (f: .31, p= .053). For Parent Investment during the second half of treatment,
Father Psychopathology was signiﬁcantly correlated (r2=-.3l, p< .05), and Mother
Psychopathology approached signiﬁcance (r2=-.29, p= .07). Further, baseline level of Positive
Child Behavior was positively correlated with Phase 2 Parent Investment ({2= .42, p< .05).
Cumulative Parent Investment from Session 1 to termination was also correlated with
pretreatment Positive Child Behavior (g2= .32, p< .05). Cumulative Investment was not
signiﬁcantly associated with baseline parent or family characteristics, although the correlation
with Father Psychopathology again approached signiﬁcance (_r_2=-.27, p= .09). Thus, parents

whose child exhibited more prosocial behaviors prior to treatment were more highly invested in

85

the intervention protocol. Level of mother and father psychopathology was not signiﬁcantly
correlated with measures of parent treatment investment, with the exception of father
psychopathology and investment during Phase 2 of the protocol.

Parent and therapist treatment satisfaction. Pretreatment level of Father Psychopathology
was negatively correlated with reported Parent Satisfaction at mid-treatment (1:2=-.37, p< .05),
termination (f: -.47, p< .01), and follow-up (:2: -.42, p< .01). Baseline Mother
Psychopathology was negatively correlated with Parent Satisfaction at mid-treatment (f: -.32,
p<.05) and termination (LE-.37, p<.05), but not signiﬁcantly so at follow-up (g2=-.28,
p= .07). Father Psychopathology was also negatively correlated with Therapist Satisfaction
ratings at both the mid-treatment (f: -.39, p< .05) and termination (g2: -.36, p< .05) time
points. Thus, higher levels of both maternal and paternal psychopathology were associated with
less treatment satisfaction and more negative expectations about subsequent progress among
participating parents. In contrast, therapists’ satisfaction with the families’ progress and
expectations for continued improvement appear to have been more exclusively tied in with level
of paternal psychopathology, despite the fact that only a subset of fathers in the study actually
participated in the intervention. This supports other evidence that fathers may make a unique
contribution to the success of behavioral interventions (Horton, 1984; Patterson & Stoolmiller,
in press), and the suggestion that fathers who are not participating in a treatment program may
nonetheless have an impact on treatment process and outcome characteristics (Nye, et al. , in
press).

Relationships Among Treatment Process Characteristics

In the next set of analyses, the relationships among pretreatment parent and therapist

expectations, parent treatment investment, and parent and therapist satisfaction ratings were

examined. It was anticipated that initial parent and therapist expectations would be positively

86

associated with subsequent parent investment in the ﬁrst half of treatment (Phase 1). In turn,
Phase 1 Parent Investment was expected to be positively associated with Parent and Therapist
Satisfaction ratings (including expectations for future change) at mid-treatment. Finally, mid-
treatrnent Parent and Therapist Satisfaction levels were expected to be related to subsequent
(Phase 2) Parent Treatment Investment, with greater satisfaction at mid-treatment associated with
higher levels of investment in the second half of treatment. These hypotheses were tested using
a series of bivariate correlations and hierarchical regressions.
Pretreatment Expectations and Subsequent Investment

Two regression equations examined parent and therapist pretreatment expectations as
predictors of subsequent parent treatment investment during Phase 1 (Early Parent Investment).
Parent Expectations were signiﬁcantly associated with Early Investment (E: .39, F(l,27)=4.7,
p<.05), whereas Therapist Expectations were not (32:09, F(l,27)= .23, p=n.s.). Thus,
parents’ initial expectations about the potential beneﬁts of the intervention were predictive of
parents’ subsequent investment in the ﬁrst half of treatment, whereas therapists’ expectations were
not related to observed parent investment. Because a signiﬁcant portion of the data involving
parent and therapist expectations at baseline was estimated from other pretreatment parent and
family characteristics, these relationships were also tested for the subset of families for whom
these data were not estimated. These analyses revealed similar trends, with Parent Expectations
associated with Early Treatment Investment (32:.19, F(l,18)=4.26, p=.054) and Therapist
Expectations showing no such relationship (32: .06, F(l,12)= .71, p=n.s.).
Treatment Investment and Parent and Therapist Satisfaction

The next set of analyses examined the extent to which Early (Phase 1) and Late (Phase
2) Parent Treatment Investment measures were associated with Parent and Therapist Satisfaction

(including expectations for future change) at mid—treatment and termination. It was anticipated

87

that parent and therapist reports of satisfaction would be positively associated with both prior and
subsequent levels of investment exhibited by the parents. Through a series of bivariate (Pearson)
correlational analyses, Parent and Therapist Satisfaction at mid-treatment and termination were
observed to be signiﬁcantly associated with Early and Late Parent Investment (Table 3).
Because of the observed association between parent treatment investment and parent and
therapist satisfaction at different points in treatment, it was considered important to next examine
whether treatment investment and satisfaction were predictive of one another with prior levels
controlled for. Thus, hierarchical regression analyses were used to examine the temporal linkages
between treatment investment and satisfaction ratings over the course of the intervention, with
the treatment process model broken down into more manageable components:
Pretreatment Phase 1 Mid-treatment Phase 2 Termination
1) Expectation —- > Investment--> Satisfaction
2) Investment-- > Satisfaction ------ > Investment
3) Satisfaction-«~- > Investment- > Satisfaction
For each equation, the predictor variable (Investment or Satisfaction) was entered on Step
2, after controlling for prior level of the dependent variable. In examining the ﬁrst proposed
relationship, between Phase 1 Parent Treatment Investment and Mid-treatment Parent Satisfaction,
Pretreatment Parent Expectations were entered ﬁrst into the equation, followed by the Early
Investment score. In this analysis, Early Parent Investment was a signiﬁcant predictor of
subsequent Parent Satisfaction, but Pretreatment Parent Expectations were not (Table 4). Thus,
parents’ level of investment and motivation demonstrated during the ﬁrst half of treatment was
positively associated with the extent to which parents reported at mid-treatment that they were
satisﬁed with the intervention thus far and that they anticipated continued success in the program.

The hypothesized predictive equation for Mid-treatment Therapist Satisfaction ratings was not run

88
Table 3
Pearson Correlations Between Early and Late Parent Treatment Investment and Parent and

Therapist Satisfaction Ratings at Mid—treatment and Termination

Parent Treatment Investment

 

 

Satisfaction Rating Early Investment Late Investment
Mid-treatment Parent Satisfaction .50“ .44**
Mid-treatment Therapist Satisfaction .71*** .59***
Termination Parent Satisfaction .50** 53*"
Termination Therapist Satisfaction .50** .60***

 

**p< .01. ***p<.001.

89

because it was considered to be redundant with the results of the earlier correlational analyses,
which already demonstrated that Phase 1 Parent Investment was signiﬁcantly associated with mid-
treatment Therapist Satisfaction (12:.71, p< .001), while Pretreatment Therapist Expectations
were not (:2: -.05, p=n.s.). Thus, satisfaction and positive expectations reported by both
parents and therapists at the mid-treatment time point were related to the level of investment
shown by participating parents during the ﬁrst half of the intervention.

For the second proposed set of hierarchical regression analyses, mid-treatment parent and
therapist satisfaction ratings were examined separately as predictors of parent treatment
investment during Phase 2, controlling for prior (Phase 1) level of parent investment. In these
analyses, neither parent nor therapist satisfaction ratings were signiﬁcantly associated with
subsequent parent treatment investment when Early Parent Investment was accounted for. Thus,
greater satisfaction by parents and therapists at mid-treatment, although signiﬁcantly correlated
with higher levels of subsequent parent treatment investment, did not contribute to the prediction
of Phase 2 Investment beyond the inﬂuence of Phase 1 Investment. In testing the ﬁnal segment
of the predictive model (Equation 3), separate regression analyses were conducted with Late
(Phase 2) Parent Investment as a predictor of subsequent Parent and Therapist Satisfaction at
termination, controlling for prior (mid-treatment) level of satisfaction. In predicting Parent
Satisfaction at termination, Late Parent Treatment Investment only approached signiﬁcance at Step
2, after mid-treatment Parent Satisfaction was entered into the equation (A32: .08, p= .058).
Similarly, Therapist Satisfaction at mid-treatment was found to be the main contributor in the
prediction of subsequent Therapist Satisfaction at termination $2 = .52, E(1 ,27) = 28. 81 , p < .001),
as Late Parent Treatment Investment did not account for a signiﬁcant portion of the variance in
Step 2 of the predictive equation (ARZ= .04, p= .11). Thus, although signiﬁcantly correlated with

level of prior parent treatment investment, both parent and therapist satisfaction ratings at

90
Table 4

Hierarchical Regression Analysis of Early (Phase 1) Parent Treatment Investment as a Predictor

of Mid-treatment Parent Satisfaction, With Pretreatment Parent Treatment Expectations Controlled

 

Treatment Process Variable B §_ B [_3 33

Step One .05
Pretreatment Parent Expectations .17 .14 .23

Step Two 25*
Pretreatment Parent Expectations .03 .14 .04
Early Treatment Investment .69 .26 .48*

 

*p< .05.

91

termination were most strongly associated with level of satisfaction with the program at mid-
treatment.

Together, these analyses demonstrate that parent treatment investment was signiﬁcantly
correlated with parent and therapist satisfaction levels across time, but neither investment nor
satisfaction was predictive of the other when prior levels of investment and satisfaction were
accounted for. One possible explanation for this is that parent investment and parent and
therapist satisfaction did not change signiﬁcantly between the time points at which they were
measured in this intervention, thereby making the earlier and later indices of these essentially the
same. This explanation may partially account for these ﬁndings, as mid-treatment and
termination ratings of parent and therapist satisfaction were not observed to be signiﬁcantly
different (M=.04, £3; .56 at mid-treatment and M=.02, §Q= .57 at termination for Parent
Satisfaction, t(28)= -.24, p=.815, two-tailed; and M=.00, _S_Q=.66 at mid-treatment and
M=.02, §Q=.67 at termination for Therapist Satisfaction, _t_(28)= .16, p=.87, two-tailed).
However, a signiﬁcant difference was found between Early M=3.63, §Q=.40) and Late
(M=3.31, S12: .75) Parent Investment scores, with an overall drop in investment observed
during Phase 2 of the program (t(28)=3.01, p< .01, two-tailed). An alternative explanation for
the lack of association between mid-treatment satisfaction ratings and subsequent parent treatment
investment may be that the observed drop in Phase 2 investment was somewhat experimentally
manipulated. That is, during Phase 2 of the intervention protocol, the therapists’ work with the
families purposely shifted from intensive weekly sessions focusing on targeted child behaviors
and parenting practices, to bi-weekly meetings to address communication and problem-solving
skills between the parents, as well as to continue to stabilize the child management techniques
developed in Phase 1. Other reports on these families indicate that this transition was sometimes

a difﬁcult one for families (e. g. Nye, 1992), as the protocol moved away from its sole parent

92

training approach, to also address what were typically longer-term and more problematic family
interaction patterns. Because the satisfaction ratings made by parents and therapists at mid-
treatrnent were based on their feelings regarding the previous child-focused work, they would not
necessarily predict parents’ subsequent reactions to and investment in the marital problem-solving
aspects of the program.
Parent and Therapist Satisfaction

Correlations between parent and therapist satisfaction ratings were also examined to
evaluate the extent of similarity in parents’ and therapists’ perceptions of treatment progress and
process. Parent and Therapist Satisfaction were highly correlated at mid-treatment (:2: .48,
p< .01, two-tailed) and termination (52: .73, p< .001, two-tailed). In addition, therapists’
expressed satisfaction with the success of the intervention was also correlated with parents’
reports of satisfaction at the six-month follow-up (f= .61, p< .001, two-tailed). Thus, parents
and therapists were found to have generally similar perceptions regarding the effectiveness of the
treatment and expectations that the program would continue to be beneﬁcial to the families (Table
5).

Relationships Among Treatment Process and Treatment Outcome Variables

The analyses conducted thus far have identiﬁed which pretreatment parent and family
characteristics were associated with subsequent treatment process and outcome, and have also
examined the relationships among treatment process characteristics. What remains to be
examined is whether aspects of the treatment process are predictive of treatment outcome, and
if this association is maintained when pretreatment predictors of outcome are controlled. In
reference to the ﬁrst question, the next set of analyses examined the association between treatment
process characteristics (parent and therapist satisfaction, and parent treatment investment) and

treatment outcome (changes in child behavior and parenting). It was anticipated that both parent

93

and therapist satisfaction ratings at Post—test would be positively associated with improvement in
child behavior and parenting from pretreatment to the same time points. Further, it was
hypothesized that level of parent treatment investment would be predictive of subsequent
parenting and child behavior ratings, with greater parent investment associated with more positive
parent and child outcomes.
Treatment Satisfaction and Child Behavior Change

In order to examine the relationship between satisfaction and child outcome, partial
correlations were generated, controlling for pretreatment level of child behavior. Parent
Satisfaction at mid-treatment, termination, and follow-up was expected to be positively correlated
with Positive Child Behavior and negatively correlated with Negative Child Behavior at the same
time points. Similarly, Therapist Satisfaction was anticipated to be positively associated with
changes in Positive Child Behavior from pretreatment to the mid-treatment and termination time
points, and to be negatively associated with Negative Child Behavior at these time points when
pretreatment Negative Behavior was partialled out. Overall, these hypotheses were supported,
with both parent and therapist satisfaction ratings correlated in the expected direction with post-
test ratings of child behavior (Table 6). The only relationships which were not found to be
signiﬁcant at the 95 % conﬁdence level were Parent Satisfaction with Positive Child Behavior at
mid-treatment (f: .21, p= .14, one—tailed) and with Negative Child Behavior at follow-up (£2:
-.30, p=.06, one-tailed).
Treatment Satisfaction and Changes in Parenting

Partial correlations were also generated in order to examine the relationship between
satisfaction and parenting outcome, with pretreatment parenting partialled out. It was anticipated
that parent and therapist reports of satisfaction with treatment would be positively associated with

concomitant changes in parenting from pretreatment to the mid-treatment and termination time

94
Table 5

Pearson Correlations between Parent and Therapist Satisfaction Ratings at Post-test Time Points

Therapist Satisfaction Rating

 

 

Parent Satisfaction Rating Mid-treatment Satisfaction Termination Satisfaction
Mid-treatment Satisfaction .48** .50**
Termination Satisfaction .53** .73****
Follow—up Satisfaction .39* 51“"

 

*p<.05,two-tailed. **p<.01,two-tailed. ***p<.001,two-tailed. ****p<.0001,two—tailed.

95

points. At mid-treatment (Post-test 1), both Parent and Therapist Satisfaction were correlated
with change in Parenting, with increases in parent authoritativeness from pretreatment to mid-
treatment coinciding with greater parent and therapist satisfaction with treatment. Similarly,
increased authoritative parenting from pretreatment to termination was associated with greater
parent and therapist satisfaction at termination (see Table 6).

2am Treatmgt Investment and Child Behavior Change

The role of parent treatment investment in accounting for changes in child behavior from
pretreatment levels was examined via hierarchical regression analyses conducted for the positive
and negative child behavior means at mid-treatment, termination, and follow-up. Changes in
child behavior from mid-treatment to termination and from termination to follow-up were not
examined, because of the earlier observation that initial improvement in child behavior at mid-
treatment remained relatively constant thereafter. In all cases, pretreatment child behavior was
entered in Step 1 of the equations, followed by parent treatment investment in Step 2.
Interactions between parent treatment investment and pretreatment child behavior were also
examined but are presented in Appendix H because of their exploratory nature and less stringent
effect size criterion.

Positive child behavior. Three regression equations were used to examine changes in
positive child behavior from pretreatment to the mid-treatment (PT 1), termination (PT 2), and
follow-up (PT 3) time points, using previous level of parent treatment investment as a predictor
of subsequent child behavior. It was anticipated that parent treatment investment during Phase
1 of the intervention program would be a positive predictor of positive child behavior at mid-
treatment, with pretreatment level of positive behavior controlled. For the termination and
follow-up time points, it was hypothesized that the cumulative level of parent treatment

investment from Session 1 to termination would be positively associated with changes in positive

Table 6

96

Partial Correlations (One-tailed) Comparing Parent and Therapist Satisfaction Ratings with Child

and Parent Outcomes at the Same Time Points, Controlling for Baseline Child Behavior and

Parenting Style

Treatment Outcome Measure

 

Positive Child Negative Child Authoritative

 

Satisfaction Rating Behavior Behavior Parenting
Parent Satisfaction
Mid-treatment .21 -.51** .53"
Termination .38* -.52** .43*
Follow-up .45** -.30 --
Therapist Satisfaction
Mid-treatment .33* -.58*** .73***
Termination .54** -.64*** .63***

 

*p< .05. **p< .01. *"‘*p< .001.

97

child behavior. These analyses indicate that Early Parent Treatment Investment was not a
signiﬁcant predictor of Positive Child Behavior at mid—treatment, with pretreatment positive
behavior controlled (AI;2= .10, p=.055). For changes in Positive Child Behavior from
pretreatment to termination, Cumulative Parent Investment was a predictor, with baseline positive
behavior controlled (A32: .31 , p< .001). However, when Positive Child Behavior at follow-up
(PT 3) was examined, Cumulative Investment was not predictive beyond the contribution of
pretreatment Positive Child Behavior. Thus, parent treatment investment was found to account
for changes in positive child behavior from pretreatment to termination, but not at the mid-
treatrnent or follow-up time points (Table 7).

Early Treatment Investment was also tested as a predictor of Positive Child Behavior at
termination. With pretreatment level of positive behavior controlled, Early Investment was not
a signiﬁcant predictor of termination level of positive child behavior (A32=.10, p= .07). Thus,
initial (Phase 1) levels of parent treatment investment did not account for change in positive child
behavior occurring during the intervention.

Negative child behavior. The inﬂuence of parent treatment investment on changes in
negative child behavior from pretreatment to the mid-treatment, termination, and follow-up time
points was also examined using similar hierarchical regression analyses. Level of parent
treatment investment during Phase 1 of the intervention program was expected to be negatively
associated with negative child behavior at mid-treatment, with pretreatment negative child
behavior partialled out. For negative child behavior at termination and follow-up, it was
anticipated that cumulative parent treatment investment across all sessions would be negatively
related to changes in negative child behavior from pretreatment level. In these analyses, Early
Parent Treatment Investment was predictive of mid-treatment (PT 1) Negative Child Behavior,

with pretreatment Negative Behavior controlled. Cumulative Treatment Investment was also

98
Table 7

Hierarchical Regression Analyses of Parent Treatment Investment as a Predictor of Positive Child

Behavior from Baseline (Pretreatment) to Mid-treatment, Termination, and Follow-up Levels

Predictor Variable B SE B B B3

 

Mid-treatment Positive Child Behavior

Step One .28**
Baseline Positive Child Behavior .55 .17 .53**

Step Two .38**
Baseline Positive Child Behavior .38 .18 .37*
Early Treatment Investment .36 .18 .35

Termination Positive Child Behavior

Step One .15*
Baseline Positive Child Behavior .41 .19 .39*

Step Two .3 1* *
Baseline Positive Child Behavior .21 .19 .20
Cumulative Treatment Investment .37 .15 .45*

Follow-up Positive Child Behavior

Step One .15*
Baseline Positive Child Behavior .43 .20 .39*

Step Two .17
Baseline Positive Child Behavior .37 .22 .33
Cumulative Treatment Investment .12 .17 .14

 

*p<.05. **p<.01. ***p<.001.

99

found to make a unique contribution in the prediction of Negative Child Behavior from
pretreatment to both termination and follow-up (Table 8).

Early Investment was also tested as a predictor of change in Negative Child Behavior at
termination, from pretreatment level. With pretreatment Negative Behavior entered ﬁrst in the
regression equation, Early Treatment Investment was signiﬁcantly associated with Negative Child
Behavior at the end of the intervention (A3}: .16, p< .05). Thus, greater investment shown by
parents in the ﬁrst half of the intervention was associated with decreased negative child behavior
approximately six months later. This suggests that therapist ratings of parent investment were
not merely correlational with changes in child behavior, but that treatment investment contributed
to observed child outcome.

Because of the observed relationship between parent pretreatment expectations and
negative child behavior at mid-treatment, an additional hierarchical regression of early investment
on subsequent negative child behavior was conducted with parent expectations controlled. In this
analysis, Parent Expectations lost their predictive association with mid-treatment Negative Child.
Behavior, suggesting that Early Treatment Investment was a mediator of pretreatment parent
expectations on subsequent child outcome (Table 9).

Parent Treatment Investment and Changes in Parenting

Parent treatment investment was also examined as a predictor of changes in level of
authoritative parenting style from pretreatment to mid-treatment and termination. It was
hypothesized that parent treatment investment during Phase 1 of the intervention (Early
Investment) would be positively associated with parenting at mid-treatment, beyond the inﬂuence
of pretreatment parenting. In addition, it was anticipated that level of parent investment
throughout the course of the intervention would be a positive predictor of change in parenting

from pretreatment to termination. Again, changes in parenting from mid-treatment to termination

100
Table 8

Hierarchical Regression Analyses of Parent Treatment Investment as a Predictor of Negative

 

Child Behavior from Baseline to Mid-treatment, Termination, and Follow-up Levels
Predictor Variable B _S_B B B R2

 

Mid-treatment Negative Child Behavior

Step One .26**
Baseline Negative Child Behavior .52 .17 .51**

Step Two .55***
Baseline Negative Child Behavior .47 .14 .46**
Early Treatment Investment -.54 .13 -.54***

Termination Negative Child Behavior

Step One .24**
Baseline Negative Child Behavior .49 .17 .49**

Step Two .55***
Baseline Negative Child Behavior .44 .13 .43**
Cumulative Treatment Investment -.45 .11 -.56***

Follow-up Negative Child Behavior

Step One .25**
Baseline Negative Child Behavior .60 .20 .50**

Step Two .39**
Baseline Negative Child Behavior .56 .18 .47**
Cumulative Treatment Investment -.34 .14 -.37*

 

*p< .05. **p< .01. ***p< .001.

101
Table 9

Hierarchical Regression Analysis of Early (Phase 1) Parent Treatment Investment as a Predictor
of Mid-treatment Negative Child Behavior, With Pretreatment Parent Treatment Exmgtations and

Negative Child Behavior Controlled

 

Predictor Variable B B_ B B B2

Step One .26**
Pretreatment Negative Behavior .52 .17 .51**

Step Two .40"
Pretreatment Negative Behavior .57 .16 .55**
Pretreatment Parent Expectations -.21 .08 -.38*

Step Three .58****
Pretreatment Negative Behavior .50 .13 .49**
Pretreatment Parent Expectations -.11 .08 -.20
Early Treatment Investment -.46 .14 -.46**

 

*p< .05. **p< .01. ***p< .001. ****p< .0001.

102

were not examined, because these post-test levels were not found to deviate signiﬁcantly from
one another. In these analyses, Authoritative Parenting at mid-treatment was positively associated
with Early Parent Treatment Investment, beyond the inﬂuence of pretreatment parenting
(ABZ=.25, p< .001). Similarly, change in Parenting from pre—treatrnent to termination was
positively associated with Cumulative Parent Investment (A32: .43, p< .0001) (Table 10).

In order to test the direction of effect between treatment investment and changes in
parenting, Early Investment was also examined as a predictor of Authoritative Parenting at
termination. This analysis identiﬁed a signiﬁcant positive relationship between Phase 1 Treatment
Investment and Authoritative Parenting at termination, with pretreatment parenting controlled
(3’: .34, p<.0001).

Parent, Family, and Treatment Process Characteristics in th_e Prediction of Treatment Outcome

Analyses conducted thus far have identiﬁed both pretreatment and treatment process
characteristics as inﬂuences on changes in child behavior and parenting from pretreatment levels.
In this regard, both father and mother psychopathology were positively associated with level of
negative child behavior at the end of treatment. In addition, father psychopathology was
negatively correlated with positive child behavior and authoritative parenting at termination.
Regarding treatment process characteristics, parent treatment investment and parent and therapist
satisfaction ratings were all positively associated with improvements in child behavior and
authoritative parenting. However, the extent to which these treatment process characteristics
are associated with treatment outcome beyond the contribution of pretreatment parent and family
characteristics remains to be examined. In light of research suggesting that treatment process
characteristics make a unique contribution to treatment outcome (e. g. Nye, et al., in press;
Patterson & Chamberlain, 1988), it was hypothesized that measures of parent investment and

parent and therapist satisfaction would be associated with changes in both child behavior and

103
Table 10
Hierarchical Regression Analyses of Parent Treatment Investment as a Predictor of Authoritative

Parenting from Baseline (Pretreatment) to Mid-treatment and Termination Levels
Predictor Variable B _S_B B B R2

 

Mid-treatment Parenting Style

Step One .21"‘
Baseline Parenting Style .55 .21 .46*

Step Two .67***
Baseline Parenting Style .29 .14 .24
Early Treatment Investment 1.00 .17 .71***

Termination Parenting Style

Step One .22**
Baseline Parenting Style .60 .21 .47**

Step Two .65****
Baseline Parenting .35 .15 .28*
Cumulative Treatment Investment .81 .14 .68****

 

*p_<.05. **p<.01. ***p<.001. ****p<.0001.

104

parenting at termination. Further, it was anticipated that this effect would be maintained when
relevant pretreatment parent and family characteristics were controlled.

In order to test this hypothesis, conceptually guided hierarchical regression analyses were
conducted to predict changes in positive child behavior, negative child behavior, and authoritative
parenting at termination, from pretreatment levels. At Step 1 was entered the relevant baseline
child or parent measure as the index of change. In Step 2 were entered what were considered
to be distal, pretreatment inﬂuences on outcome. In this regard, both mother and father
psychopathology were identiﬁed earlier as pretreatment correlates of child behavior and parenting
outcomes. Because Mother Psychopathology was observed to have substantial but not always
signiﬁcant association with the outcome measures, a combined Parent Psychopathology variable
was used to maximize the potential contribution of pretreatment parent functioning on child and
parent outcome, as well as to reduce the number of variables considered. The Family Resources
variable was not included as a predictor because the earlier partial correlation analyses did not
determine it to be signiﬁcantly related to any of the outcome measures at termination. In Step
3 of the equations were entered the more proximal, treatment-related predictors of outcome at
termination. In order to reduce the number of variables in the equations, and because of the
substantial colinearity observed among these treatment process constructs, investment and
satisfaction ratings were combined into one "Treatment Process" variable comprised of
cumulative parent and therapist satisfaction ratings and parent treatment investment means
(Alpha=.83). This variable was entered as the ﬁnal predictor in the regression equations,
representing the most proximal predictor of child and parenting outcomes.

Cﬂpge in Positive and Negative Child Behavior at Termination
The equation for Positive Child Behavior at termination included pretreatment Positive

Child Behavior, Parent Psychopathology, and the Treatment Process construct. Treatment

105

Process was not found to be a signiﬁcant predictor of Post-test 2 Positive Child Behavior with
pretreatment Positive Behavior and Parent Psychopathology controlled (Table 11). For Negative
Child Behavior at termination, pretreatment Negative Child Behavior (Step 1), and Parent
Psychopathology (Step 2) were entered into the equation prior to the Treatment Process cluster.
In this analysis, Treatment Process characteristics were negatively associated with Negative Child
Behavior at termination (A32: .26, p< .001), but Parent Psychopathology was not a signiﬁcant
predictor (Table 12). Thus, greater parent investment and parent and therapist satisfaction was
the primary predictor of decreased negative child behavior at termination, from pretreatment
level.
Change in Authoritative Parenting at Termination

The predictive equation for Post-test 2 Parenting included pretreatment Parenting (Step
1) and Parent Psychopathology (Step 2), in addition to the Treatment Process variable. The
Treatment Process measure was positively associated with Parenting at termination (AB2=.16,
p< .001) with pretreatment parenting and parent psychopathology controlled (Table 13),
indicating that greater parent investment and parent and therapist satisfaction were related to more
authoritative parenting at the end of treatment.
Tests of Mediatﬂ

In light of the observed correlation of mother and father psychopathology with parent
treatment investment and parent and therapist satisfaction ratings, the ﬁnal analyses examined the
extent to which the Treatment Process construct served as a mediator of the impact of parent
psychopathology on child and parent outcome. In order to demonstrate that a variable is a
mediator between an independent and a dependent variable, the independent variable and the
presumed mediator must be related both to one another and to the dependent variable (Baron &

Kenney, 1986). Therefore, partial correlations associating the Parent Psychopathology and

106
Table 11

Hierarchical Regression Analysis of Parent Psvchopathology and Treatment Process
Characteristics as Predictors of Positive Child Behavior from Baseline (Pretreatment) to

Termination Level

 

Predictor Variable B S_ B B _R_2
Step One .15*
Baseline Positive Child Behavior .41 .19 .39*
Step Two .33**
Baseline Positive Child Behavior .44 .17 .42*
Parent Psychopathology -.08 .03 -.42*
Step Three .40"
Baseline Positive Child Behavior .35 .17 .33
Parent Psychopathology ~.05 .03 «.27
Cumulative Treatment Investment .22 .13 .32

 

*p<.05. **p<.01. ***p<.001.

107
Table 12

Hierarchical Regression Analysis of Parent Psychopathology and Treatment Process

Characteristics as Predictors of Negative Child Behavior from Baseline (Pretreatment) to

Termination Levels

 

Predictor Variable B ﬂ?) B B B2
Step One .24**
Baseline Negative Child Behavior .49 .17 .49**
Step Two .31**
Baseline Negative Child Behavior .42 .17 .41*
Parent Psychopathology .05 .03 .28
Step Three .57***
Baseline Negative Child Behavior .51 .14 .50**
Parent Psychopathology -.002 .03 -.01
Treatment Process Characteristics -.38 .10 -.58***

 

*p< .05. **p< .01. ***p< .001.

108
Table 13

Hierarchical Regression Analysis of Parent Psychopathology and Treatment Process
Characteristics as Predictors of Authoritative Parenting Sgle from Baseline (ﬂetreatment) to

Termination

 

 

Predictor Variable B BB B B Bf

Step One .22M
Baseline Parenting Style .60 .21 .47**

Step Two .34”
Baseline Parenting Style .62 .20 .49**
Parent Psychopathology -.09 .04 -.34*

Step Three .60****
Baseline Parenting Style .54 .16 .43”
Parent Psychopathology -.01 .04 -.04
Treatment Process Characteristics .76 .19 .59***

 

*p<.05. **p_<.01. ***p_<.OOl. ****p<.0001.

109

Treatment Process constructs with termination levels of child behavior and authoritative parenting
were generated. These indicate signiﬁcant relationships between Parent Psychopathology and
both Positive Child Behavior ([2: -.46, p< .01, one-tailed) and Authoritative Parenting ([2: -
.38, p < .05, one-tailed) at termination, controlling for baseline child and parent functioning. The
Treatment Process construct was associated with termination levels of Positive Child Behavior
(:2: .52, p<.01, one-tailed), Negative Child Behavior (:2: -.57, p<.001, one-tailed), and
Authoritative Parenting (£2= .68, p < .001, one-tailed), with pretreatment levels of these outcome
measures controlled. Finally, Parent Psychopathology and Treatment Process Characteristics
were observed to be signiﬁcantly associated with one another, based on a one-tailed test of their
bivariate (Pearson) correlation (g2: -.47, p< .01). Therefore, the initial requirements for
demonstrating mediation were met for the Positive Child Behavior and Authoritative Parenting
outcome measures. However, only the Authoritative Parenting equation was tested for mediation
because of the observation that Treatment Process Characteristics did not directly predict Positive
Child Behavior at termination beyond the contribution of baseline child behavior and parent
psychopathology.

In order to test the hypothesis that treatment process characteristics mediated the
relationship between parent psychopathology and parenting outcome, a stepwise regression
analysis was conducted. If treatment process characteristics mediate the effects of parent
psychopathology on parenting outcome at termination, the Treatment Process cluster would be
expected to enter into the predictive equation, while the Parent Psychopathology index would
presumably be omitted. Therefore, the predictive equation was set up to ﬁrst enter pretreatment
level of authoritative parenting as the index of change, and then to consider Parent
Psychopathology and Treatment Process Characteristics in a stepwise fashion on Step 2.

In predicting change in Authoritative Parenting at termination from pretreatment level, Treatment

110

Process characteristics emerged as a salient predictor (A32: .38, p< .001), whereas Parent
Psychopathology did not. Thus, treatment process characteristics appear to account for the
relationship between pretreatment parent psychopathology and termination level of authoritative
parenting within this sample of treatment families.
Power Analyses

In order to evaluate the magnitude of the observed effect sizes for this relatively small
treatment sample, statistical power analyses were conducted on the ﬁnal equations used in
predicting treatment outcome. Regarding the overall predictive equations for positive and
negative child behavior and authoritative parenting, power ranged from 82 to 99 at p< .01,
suggesting that the predictive ability of these relationships is signiﬁcantly better than what would
be expected by chance. Power was also substantial regarding the unique contribution made by
the treatment process characteristics in predicting negative child behavior and authoritative
parenting at termination (estimated at 81 and 75, respectively, at p< .05). Power for the test of
mediation between parent psychopathology and treatment process characteristics was also
determined to be adequate (approximately 99 overall, and 87 for Treatment Process

Characteristics, at p< .01).

CHAPTER IV
DISCUSSION

This study provides substantial support for the role of treatment process characteristics
as proximal determinants of treatment outcome in parent training interventions, with pretreatment
parent, family, and child characteristics representing more distal inﬂuences on treatment success.
In particular, these ﬁndings indicate that parent investment, parent and therapist expectations, and
parent and therapist satisfaction are important components of treatment that were directly
associated with changes in child behavior and parenting style among the subset of families who
completed the early intervention parent training protocol. This is consistent with a small but
growing body of literature which has also identiﬁed parent expectations (Clark & Baker, 1983),
parent satisfaction (Furey & Basili, 1988), and parent treatment investment (e. g. Nye, et al., in
press; Patterson & Chamberlain, 1994; Patterson, Dishion, & Chamberlain, 1993; Patterson &
Stoolmiller, Stoolmiller, et al., 1993) as correlates of outcome in parent training interventions.
The present work expanded on this previous research by including therapist perceptions of
treatment as potential inﬂuences on outcome, and by examining the interplay among these various
parent and therapist treatment process characteristics as predictors of treatment effectiveness. In
addition, this report represents both a reﬁnement and a substantial extension upon earlier work
with families in the M.S.U. program, through the inclusion of paternal as well as maternal
investment ratings, the use of multiple reports of child behavior, the examination of authoritative
parenting style as an additional outcome variable, and the focus on parent and therapist treatment
expectations and satisfaction as other process characteristics associated with treatment success.
Pretreatment Inﬂuences on Outcome

Overall, the M.S.U. intervention program was found to be successful in effecting change

in child and parenting outcome from pretreatment levels, as comparisons of the child behavior

lll

112

means and authoritative parenting scores across pretest and post-tests revealed increased positive
child behavior, decreased negative child behavior, and more authoritative parenting from
baseline. Further, as expected, changes in child behavior were correlated with changes in
parenting at the mid-treatment and termination time points, but were not associated with
pretreatment parenting style. The ﬁnding that the intervention contributed to both a reduction
in negative child behavior and an increase in positive child behavior is particularly signiﬁcant in
light of evidence that preschool-aged children with conduct problems not only exhibit an extreme
rate of aversive behaviors, but also display fewer prosocial interactions (Gardner, 1987).
Therefore, improved outcomes on both of these dimensions would be expected to enhance the
ability of this intervention program to prevent the development of child conduct problems later
on.

Subsequent analyses identiﬁed which pretreatment parent and family characteristics
represented distal inﬂuences on child and parenting outcomes. Of particular importance is the
observation that father psychopathology was the one pretreatment characteristic that was
consistently associated with positive child behavior, negative child behavior, and authoritative
parenting at the various post-test time points. The association between fathers’ functioning and
subsequent child behavior is consistent with other research showing that fathers play a signiﬁcant
role in the development of childhood psychopathology (e. g. Phares & Compas, 1992). Further,
the negative correlation between father psychopathology and authoritative parenting indicates that
both fathers’ and mothers’ acquisition of new parenting techniques may have been impaired by
the level of paternal distress. This supports other research which has found an association
between paternal depression and impaired maternal child-rearing practices (Davies, et al., 1989;
Schneider, et al., 1989). Although mother psychopathology was not observed to be as

consistently correlated with changes in positive child behavior, it should be noted that this

113

treatment sample was selected because of the fathers’ alcohol problems, with the expectation that
sons of these alcoholic fathers would be at signiﬁcantly elevated risk for the development of
problem behaviors in childhood and problems with alcohol later in life. Therefore, level of father
psychopathology in this sample was anticipated to be a particularly salient correlate of child
behavior. This is further supported by research on the larger group of families in the longitudinal
parent project, which found a signiﬁcant positive association between paternal psychopathology
and problematic child symptomatology (Jansen, Fitzgerald, Ham, & Zucker, in press).

In addition to their role as distal predictors of child and parenting outcome, pretreatment
parent, child, and family characteristics were observed to be indirectly correlated with treatment
outcome through their association with the more proximal treatment process inﬂuences.
Regarding baseline treatment expectations, level of maternal psychopathology at baseline was
negatively correlated with parents’ expectations about the potential success of the program,
indicating that parents in families with more troubled mothers expected the intervention to be less
successful. The lack of association between father psychopathology and parent expectations may
be related to the fact that all of the treatment families were families in which the fathers were
known to be troubled, but within which the mothers’ level of functioning was more variable.
Thus, some families had two parents who were highly troubled, whereas others had a mother who
was relatively intact. Therefore, the correlation of mother but not father psychopathology with
pretreatment parent expectations does not necessarily mean that father psychopathology was
irrelevant, but may instead be a reﬂection of a greater level of difﬁculty among families in which
both the mother and the father experienced signiﬁcant psychological problems. This explanation
would be consistent with other research (e. g. Merikangas, et al., 1985; Williams, 1987), which
has identiﬁed an additive if not synergistic effect on family functioning when both parents are

psychologically troubled.

114

Parent expectations at baseline were also associated with the initial frequency of positive
child behaviors, indicating that parents whose child exhibited positive behaviors at a higher
frequency were more optimistic and hopeful about the potential success of the intervention for
addressing child behavior problems. This expectation is consistent with research showing that
initial levels of child compliance are positively associated with subsequent success in a parent
training program (Holden, Lavigne, & Cameron, 1990). Interestingly, neither child behavior nor
parenting style at baseline was correlated with therapist pretreatment expectations, despite the fact
that the intervention was designed to target both child and parenting behaviors for change.
Instead, therapist expectations were positively associated with family socioeconomic resources,
indicating that therapists’ impressions of families’ ability to beneﬁt from the intervention were
based primarily on the ﬁnancial, educational, and occupational status of the parents. This
expectation on the part of therapists is also supported by the parent training literature, wherein
family socioeconomic characteristics have been associated with completion of and speed of
progress in treatment (Holden, et al., 1990), with level of positive change in parenting practices
(Dumas & Wahler, 1983), and with child outcome (Webster-Stratton, 1985) in parent training
interventions. In addition, it should be noted that family resources were positively associated
with mid-treatment level of prosocial child behavior in the current treatment sample.

Both mother and father psychopathology were observed to be associated with parents’
reported satisfaction with the intervention program and expectations for continued progress.
Thus, higher levels of both maternal and paternal psychopathology were associated with less
treatment satisfaction and more negative expectations about subsequent progress among
participating parents. In contrast, therapists’ satisfaction and expectations were observed to be
more exclusively tied in with level of paternal psychopathology, despite the fact that only a subset

of fathers in the study actually participated in the intervention. In examining pretreatment

115

characteristics associated with parent investment, mother and father psychopathology were not
found to be signiﬁcantly related to early or cumulative levels, although the correlations
approached signiﬁcance. Early parent investment was somewhat negatively related to level of
authoritative parenting, which is consistent with research showing that impaired parenting at
pretreatment is associated with higher levels of maternal resistance (Stoolmiller, Duncan, Bank,
& Patterson, 1993). Parent investment at all time points was also positively correlated with
baseline positive child behavior, indicating that parents of children who were initially more
prosocial in nature were more highly invested in the intervention. Parent investment during the
second half of the intervention was signiﬁcantly correlated with father psychopathology,
suggesting that paternal distress had a disruptive inﬂuence during Phase 2 of the intervention.
This ﬁnding is particularly important in light of the hypothesis that the observed decrease in
parent investment during the second half of the protocol was related to the greater focus on
problems in individual parent and marital functioning. That level of father psychopathology was
associated with Phase 2 investment provides further support for this explanation, as families in
which the father was more troubled (e. g. more antisocial, with more severe alcoholism, and/or
more depressed) would be expected to have been more resistant to exploring and addressing
personal and marital problems.
Interplay Among Treatmen_t Process Characteristics and Treatment Outcomes

The examination of treatment expectations, investment, and satisfaction in relation to one
another and to treatment outcome provided substantial support for the hypothesized role of these
characteristics as interrelated aspects of the treatment process which inﬂuence treatment success.
Parent expectations at pretreatment were strongly associated with negative child behavior at mid-
treatrnent. In contrast, no relationship was found between pretreatment therapist expectations and

the outcome measures, although the correlation with authoritative parenting approached

116

signiﬁcance. Thus, parents’ expectations were hypothesized to have had a more signiﬁcant
inﬂuence on the early treatment process than did therapists’. This hypothesis was subsequently
supported by the ﬁnding that pretreatment parent expectations were associated with early parent
treatment investment, whereas therapist expectations were not. Further, parent treatment
investment was observed to mediate the relationship between pretreatment expectations and
negative child behavior, indicating that parent expectations affected child outcome at mid-
treatrnent indirectly, via their inﬂuence on parent investment. This also indicates that parent
expectations prior to treatment did not in and of themselves predetermine parents’ success in the
intervention.

Together, these ﬁndings are consistent with other reports demonstrating that clients’ early
attitudes about a treatment program are associated with certain aspects of treatment outcome
(Bloch, et al., 1976; Bowden, et al., 1980; Collins & Hyer, 1986), and that this association can
be accounted for by the impact of pretreatment expectations on subsequent investment in an
intervention (Bandura, 1977, 1989; Burling, et al., 1989; Elliott, 1986; Kazdin, 1981; Wolf,
1978). However, the present data are at odds with earlier research that has demonstrated a
relationship between therapist pretreatment expectations and subsequent treatment process and
outcome characteristics (Baekeland & Lundwall, 1975; Duckro, 1991; Martin & Sterne, 1975);
therapist expectations in the present study were found to be only marginally associated with
change in parenting at mid-treatment, and essentially unrelated to early parent investment, or to
parent or therapist satisfaction ratings. This disparity in ﬁndings may be attributable to the fact
that some of the other research examined therapist expectations as a predictor of premature
treatment termination. In the current work, pretreatment expectations were not examined among
treatment drop-out families; only families who completed the protocol were included in analyses.

Further, these ﬁndings involving parent and therapist expectations must be viewed as extremely

117

tentative because of the amount of missing data that needed to be imputed. However, the
observation of similar trends when these relationships were also examined within the subset of
data which was not estimated lends additional support to the validity of these results.

As anticipated, parent satisfaction and expectations at mid-treatment, termination, and
follow-up were signiﬁcantly associated with one another, indicating that parents who were more
satisﬁed with treatment were also more hopeful that they would experience continued
improvement as a result of the intervention. Similarly, therapists who were more satisﬁed with
families’ progress at mid-treatment and termination expected that observed changes would
continue over time. An examination of the roles of parent and therapist satisfaction as treatment
process characteristics indicated that both parent and therapist satisfaction ratings, which included
expectations that the intervention would continue to be effective, were signiﬁcantly correlated
with one another, and with changes in positive child behavior, negative child behavior, and
authoritative parenting at most time points. Thus, parents and therapists who reported higher
levels of overall satisfaction with the treatment generally perceived more frequent positive
behavior and decreased negative behavior in the target child, as well as a more authoritative
parenting style among participating mothers and fathers.

Parent and therapist satisfaction ratings were also associated with level of parent treatment
investment. However, neither satisfaction nor investment were predictive of one another with
earlier levels of these indices controlled. Two possible explanations for this are offered. One
is that parent investment and parent and therapist satisfaction did not change signiﬁcantly between
the time points at which they were measured, thereby making the earlier and later indices of these
essentially redundant. However, this was found to be true only for the satisfaction ratings, which
were consistent from mid-treatment to termination. In contrast, a signiﬁcant difference was found

between Phase 1 and Phase 2 parent investment, with investment dropping overall during Phase

118

2 of the program. An alternative explanation for the failure of mid-treatment satisfaction ratings
to predict parent treatment investment is that the decrease in Phase 2 investment was more
attributable to pre-determined, protocol-imposed changes in the intervention program than to
previous parent investment. In this regard, Phase 2 of the protocol involved a shift in the nature
and frequency of therapeutic contacts with the families, as the intervention moved from intensive
weekly sessions primarily focusing on targeted child behaviors and parenting practices, to bi-
weekly meetings which also targeted parent communication and problem-solving patterns.
Anecdotal accounts from therapists and supervisors involved in the program suggest that this
transition was sometimes a difﬁcult one for families (e. g Nye, 1992), because the protocol
purposely deviated from its exclusive parent training approach to also address what were often
more chronic, problematic, and affectively charged marital/family issues. Because the satisfaction
ratings made by parents and therapists at mid-treatment were based on their feelings regarding
the previous child-focused work, these ratings would not then necessarily be directly predictive
of parents’ subsequent reactions to and investment in the marital problem-solving component
which followed. Thus, the time points at which satisfaction ratings were generated may have
precluded an accurate account of variations in satisfaction during the second phase of the
intervention. In fact, there is evidence that treatment satisfaction may actually progress in a
curvilinear trend during successful interventions (Tracey, 1989). Within this framework,
satisfaction drops during the middle of treatment as more difﬁcult issues are addressed, and rises
again by termination once these issues have been successfully processed. These possible
ﬂuctuations could not be observed in the current work, as satisfaction ratings were not generated
during the course of the marital problem-solving phase.

Although not identiﬁed as a unique predictor of changes in parent and therapist

satisfaction ratings from mid-treatment to termination, parent treatment investment was predictive

119

of satisfaction at the mid-treatment time point. Further, this association between parent
investment and subsequent satisfaction was maintained when pretreatment parent expectations
were controlled, indicating that parent investment was directly related to mid—treatment
satisfaction. This is consistent with the association between investment and outcome, whereby
more positive outcomes were observed to be predicted by higher levels of investment in the
intervention. In this regard, changes in negative child behavior and authoritative parenting were
consistently associated with parent treatment investment, such that higher levels of investment
were predictive of greater improvement in negative behavior and parenting. Further, the ﬁnding
that early investment was also a signiﬁcant predictor of later levels of negative child behavior and
authoritative parenting suggests that parent treatment investment represents a true inﬂuence on
treatment outcome, rather than a result of therapists attributing greater investment to families
showing improvement during treatment. Increased positive child behavior at termination was also
associated with greater prior parent investment, when pretreatment positive behavior was
controlled. However, level of parent investment did not predict positive child behavior at mid-
treatrnent or follow-up, suggesting that other mediating mechanisms may have come into play at
these time points.

Overall, then, parent treatment investment was identiﬁed as a signiﬁcant predictor of
improvement in child behavior and parenting. Further, the strong association between parent
investment early in treatment with improvement in negative child behavior and authoritative
parenting at later time points indicates that the inﬂuence of investment on outcome is quite robust.
In addition, exploratory analyses identiﬁed parent investment as a moderator of the inﬂuence of
pretreatment child and parenting characteristics on subsequent outcomes, with highly invested
families demonstrating treatment gains that were relatively independent of pretreatment level of

child behavior and parenting (see Appendix H).

120

Treatment process characteristics as proximal inﬂuences on outcome. The literature on

parent investment in parent training interventions has shown that investment is not only associated
with treatment outcome, but that this characteristic of the treatment process is directly predictive
of changes in child behavior (e. g. Nye, et al., in press) and parenting (e. g. Patterson &
Stoolmiller, in press) when pretreatment parent and family inﬂuences on outcome are controlled.
When considered as a single construct representing treatment process characteristics, treatment
investment, expectations, and satisfaction were found to predict changes in both negative child
behavior and authoritative parenting at termination, after the contributions of pretreatment level
of child or parent functioning and baseline parent psychopathology were taken into account.
Further, the inﬂuence of parent psychopathology on outcome was mediated by treatment process
characteristics, whereby the treatment process construct was observed to account for the
relationship between parent psychopathology and parenting outcome. Thus, treatment process
characteristics may be viewed as directly inﬂuencing treatment outcome.

The failure of treatment process characteristics to predict positive child behavior at
termination beyond the inﬂuence of pretreatment child behavior and parent psychopathology may
be associated with parents’ relatively greater investment in reducing antisocial child behavior,
than on enhancing prosocial behaviors during the ﬁrst half of the intervention. As noted earlier,
parents who were more invested in the protocol were those whose child exhibited more positive
behaviors at baseline. Presumably, then, these parents would be less concerned with further
increasing positive behaviors than on reducing targeted negative behaviors. In fact, an
examination of the partial correlations of investment and satisfaction ratings with child outcome
indicates stronger associations with changes in negative child behavior at mid-treatment. In
contrast, only therapist satisfaction was signiﬁcantly correlated with mid-treatment positive child

behavior, suggesting that parents were generally more concerned with the reduction of problem

121

behaviors than the promotion of positive behaviors earlier on. However, cumulative parent
investment and termination parent and therapist satisfaction ratings were all associated with
positive child behavior at the end of the intervention, suggesting relatively greater attention to
prosocial behaviors by parents during the latter half of the intervention. Therefore, the inclusion
of both early and late satisfaction and investment ratings in the treatment process construct may
have masked a possibly greater inﬂuence of Phase 2 process characteristics on prosocial positive
behavior at termination.

Implications for the treatment process model. The analyses conducted herein were
designed to test selected components of a conceptual treatment process model (see Figure 1).
This model proposes a chain of inﬂuence among parent and therapist expectations, parent
treatment investment, and parent and therapist satisfaction as determinants of treatment outcomes,
with pretreatment parent and family characteristics identiﬁed as more distal and indirect
predictors of changes in child behavior and parenting. These analyses generally conﬁrm the
expectations that treatment process characteristics would be directly associated with changes in
positive child behavior, negative child behavior, and authoritative parenting among families
completing the M.S.U. Multiple Risk Child Outreach Program. These analyses also generally
conﬁrm the hypothesized associations among parent and therapist expectations, parent and
therapist satisfaction, and parent treatment investment, as parent expectations were observed to
affect outcome via their inﬂuence on parent investment, and parent satisfaction and expectations
for continued success in treatment were found to be positively associated with one another and
with level of parent treatment investment. Although changes in treatment investment and
satisfaction ratings were not signiﬁcantly predictive of one another during the second half of the
intervention, it is hypothesized that changes in the intervention protocol during Phase 2 may have

precluded an accurate account of the interplay between parent satisfaction and investment.

122

Findings also lend support to the proposed relative contributions of pretreatment
parent and family factors and treatment process characteristics in accounting for changes in child
behavior and authoritative parenting. Among the current sample of families, treatment process
characteristics made a unique contribution in the prediction of improved negative child behavior
and authoritative parenting at termination, beyond the inﬂuence of pretreatment parent
psychopathology. In fact, parent psychopathology was not signiﬁcantly associated with
termination level of negative child behavior. Further, treatment process characteristics accounted
for the relationship between parent psychopathology and termination authoritative parenting, as
baseline level of parent psychopathology was observed to have only an indirect impact on change
in authoritative parenting at the end of the intervention, through an inﬂuence on treatment process
characteristics. Thus, level of parent trouble impaired improvement in parenting by disrupting
the treatment process.

Overall, then, this study supports a model of treatment wherein parent treatment
investment, parent and therapist expectations, and parent and therapist satisfaction are interrelated
aspects of the treatment process which account for variability in outcomes among parents
participating in this early intervention program. In contrast, pretreatment characteristics such as
parent psychopathology and family socioeconomic resources are more distal predictors of
outcome which may, in part, inﬂuence outcome through their disruption of the treatment process.

Characterizing the ﬂow of treatment. In integrating these various ﬁndings to characterize
the interplay among pretreatment and treatment process inﬂuences within the current intervention,
one must ﬁrst begin with parent pretreatment expectations, which were associated with level of
mother psychopathology and with baseline frequency of positive child behaviors. These initial
parent expectations subsequently inﬂuenced the extent to which parents initially invested in the

intervention protocol, with lower expectations associated with less investment. However,

123

pretreatment expectations did not directly account for changes in child behavior during the
protocol. Instead, parent treatment investment predicted decreased negative child behavior and
increased authoritative parenting at mid-treatment. Further, these improvements in child behavior
and parenting were positively associated with satisfaction and positive expectations expressed by
parents and therapists at mid-treatment.

Regarding processes occurring during the second half of the intervention, an overall drop
in parent treatment investment was observed from pretreatment to termination, while parent and
therapist satisfaction levels remained consistent across time. Baseline level of father
psychopathology was implicated as a disrupter of parent investment during this second phase of
treatment, indicating that families in which the father was more disturbed were less likely to
demonstrate adequate investment during the second phase of treatment. Thus, the second half
of treatment, in which salient marital and individual problems were addressed, appears to have
been particularly critical for successful child and parenting outcomes by the end of the
intervention. In accounting for overall change from pretreatment to termination, then, treatment
process characteristics may legitimately be viewed as the more proximal contributors to observed
improvement in negative child behavior and authoritative parenting, with pretreatment
characteristics such as parent psychopathology affecting outcome through their inﬂuence on the
treatment process. In contrast, treatment process characteristics did not seem to have the same
level of inﬂuence on positive child behavior.

Unique Aspects of the Study

The parent training literature has focused primarily on clinically accessed populations,
in which participating families have actively sought help in addressing problems of antisociality
and aggressiveness in their children. Thus, identiﬁcation of these samples has typically been

determined by characteristics of the children rather than of the parents. In contrast, the families

124

involved in the current study were selected because of evidence of alcoholism in the fathers, and
were speciﬁcally identiﬁed and recruited through court records of DWI arrests. Therefore, these
families were reached as part of a community-based outreach protocol identifying children at
particularly high-risk for behavior problems, and are therefore anticipated to represent a more
population-based sample rather than one that is self-selected by virtue of seeking outside
intervention for problem child behaviors.

Although the intervention was implemented with a non-clinically generated sample of
children, there is ample evidence that participating families were particularly dense in child risk
for later alcohol and drug problems, by virtue of a high rate of antisociality and multiple parent
and family stressors (Fitzgerald, Zucker, & Yang, 1995; Zucker, Fitzgerald, & Moses, 1995;
Zucker, et al., 1988). Reports on families in the larger parent project have documented
substantial ﬁnancial and social disadvantage, signiﬁcant marital conﬂict and aggression, and high
rates of parent distress including depression, antisociality, and other drug abuse within this
community-based sample. In fact, this subpopulation of alcoholic men has been shown to have
more antisocial comorbidity than the general population of alcoholics, as indicated in part by their
drunk driving convictions. Moreover, other analyses have shown that approximately 60% of
these men may be classiﬁed as Type II (antisocial, early onset) alcoholics under Cloninger’s
(1987) typology (Nye, et al., in press; Zucker, Ellis, & Fitzgerald, 1994) Thus, the group of
alcoholic men recruited through the outreach protocol is most representative of the subset of
alcoholics who are the most damaged, with the most antisocial comorbidity, the earliest onset of
alcohol dependence, and with the greatest likelihood that substance abuse will develop among
their offspring (Zucker, Fitzgerald, & Moses, 1995). Further, although these families were not
explicitly recruited because of level of child behavior problems, a substantial subset of target

children in the larger parent project have been observed to display a cumulation of internalizing

125

and externalizing behaviors that is signiﬁcantly elevated relative to other children their age
(Jansen, Fitzgerald, Ham, Zucker, in press).

The present study is also somewhat unique with regard to the scope and underlying
orientation of the intervention protocol. Different from more traditional parent training
approaches (e.g. Dumas & Wahler, 1983; Webster-Stratton, 1984), this program was modeled
after Patterson’s Social Learning Therapy (e.g. Patterson, et al. , 1982) and incorporated a marital
problem-solving component in the treatment regimen rather than adhering to a strict child-focused
paradigm. Further, in contrast to work with clinically-identiﬁed populations of children, the
focus of the present work was purposely on ways to enhance parenting techniques and child-
rearing practices in order to prevent later child behavior problems. Thus, although sessions did
address parents’ concerns about speciﬁc child behaviors, an equally visible goal was to assist
parents in learning how to manage these early problem behaviors in the hopes of reducing the
risk that they would develop into more severe difﬁculties over time. The fact that both child
behavior and parenting style were observed to change during the course of the intervention
provides initial support for this preventative approach.

Methodological Limitations

Although this research provides evidence that such process characteristics as treatment
investment, expectations, and satisfaction are important in accounting for variability in outcome
among participants in a parent training intervention, several methodological issues mandate that
these results be interpreted with some caution. The small sample size, although typical for
treatment-based studies, not only precluded a direct test of the full treatment process model, but
also limits the generalizability of the ﬁndings. Further, this limitation is compounded by the
number of analyses conducted, which reduces the overall power of the results. However, it

should be noted that the intent and effect of these analyses was generally to maximize the

126

potential inﬂuence of non-treatment process variables, prior to testing treatment process
characteristics as predictors of outcome. Further, the observation that power was generally
within the desired parameters (Cohen, 1987; 1992) despite the limited number of participants
suggests that the ﬁndings reﬂect real rather than random relationships.

Another factor limiting the generalizability of these ﬁndings is the restricted sample
studied herein. The treatment sample was quite homogenous with regard to general demographic
characteristics, as well as the speciﬁc inclusion criteria which limited the sample to Caucasian,
intact families with an alcoholic father and a preschool-aged male target child. Given the narrow
range of variability among these families, it is not possible to speculate whether or how the
observed trends in the current work might play out differently within other treatment samples.
In addition, the current sample is further limited by self—selection, representing only those
families who not only agreed to participate in the intervention protocol, but who also completed
it. As such, they may represent a more intact and less troubled sample of families with
potentially fewer personal, marital, or other difﬁculties. In fact, other analyses identiﬁed a
higher level of marital aggression among families within the larger treatment sample (r_1=104)
who terminated their participation in post-test data collection (Nye, 1992). However, even within
this more selective sample, it is noteworthy that such variability in outcome was systematically
related to treatment process characteristics.

Also worthy of comment in the present work is the cognitive interdependence of the
treatment-related variables that were studied. Parent treatment expectations and satisfaction, and
parent treatment investment cannot be viewed as mutually exclusive constructs which necessarily
represent unique aspects of the treatment process. Similarly, therapist treatment expectations and
satisfaction are also not independent of one another, nor separate from therapists’ ratings of

parent treatment investment. More likely, these process characteristics are intimately connected

127

with one another, and with perceived changes in the outcome measures. Thus, it is not possible
to conclude that any one of these characteristics necessarily drives or determines any other
process or outcome aspect of the treatment, despite the longitudinal relationships that have been
described.

Another methodological concern in this work is a potential criterion contamination bias,
whereby parents’ and therapists’ ratings of outcome may have been inﬂuenced by their
experiences of the treatment process. For example, parents who initially had more optimistic
expectations about the intervention may have reported more positive outcomes, simply as a
conﬁrmation/afﬁrmation of those expectations. However, it should be noted that it was level of
parent treatment investment - that was rated by therapists - rather than prior parent expectations
that was found to be predictive of child and parent outcome. Another possible bias may be that
parents who were more invested in treatment were then more motivated to perceive change in the
targeted behaviors, and therefore reported more positive outcomes than did less invested parents.
Similarly, therapists also may have been more likely to report improved child behavior and
parenting in families where the parents were perceived to be highly motivated and invested.
Conversely, therapists might have attributed greater investment for families in which parenting
style and child behavior appeared to be improving. ‘One way to address this latter issue was to
examine the extent to which early ratings of parent investment were associated with child and
parenting outcomes at termination. Presumably, these earlier ratings would be more independent
of perceived changes in child behavior at the later time point than at the more proximal one. The
ﬁnding that parent investment scores for the ﬁrst half of treatment were also signiﬁcantly
associated with changesin negative child behavior and authoritative parenting at termination
indicates that this particular confound may not have been a problem. In further support of this

is the ﬁnding that therapist ratings of parent investment over the course of treatment were

128

signiﬁcantly associated with level of negative child behavior at the six-month follow-up. Some
safeguards against other potential biases included the use of an average child behavior score based
on ratings from multiple sources, the validation of initial therapist ratings of parenting style
against independent observations, and reliability checks of therapists’ investment ratings.
However, the potential and actual biases involved when relying on treatment participants’ ratings
of outcome have been well—documented (Furey & Basili, 1988; Johnson & Christensen, 1975;
Patterson, Cobb, & Ray, 1973; Patterson & Chamberlain, 1988; Patterson, et al., 1993; Weist,
Ollendick, & Finney, 1991), and the most stringent safeguard against these potential criterion
contamination biases would clearly be to incorporate more objective longitudinal ratings of parent
treatment investment, child behavior, and parenting.

The issue of degree of change in child behavior and parenting must also be considered.
Treatment outcome research has begun to focus on the importance of assessing the extent to
which targeted behavior change is clinically signiﬁcant, and not merely statistically signiﬁcant
(Jacobson & Truax, 1991; Speer, 1992). Under this model, the success of interventions would
not be measured solely by examining relative change in behavior from baseline levels, but would
also be determined by the number of participants whose symptoms are reduced to a normal, non-
clinical level as a result of treatment. However, unlike most parent training studies, children in
the present work were not recruited for the intervention because of a high level of problematic
behavior. Instead, these children were identiﬁed for early intervention services because they
were considered to be at-risk for the subsequent development of problem behaviors by virtue of
their membership in an alcoholic family. Therefore, level of clinically signiﬁcant change is not
yet relevant in this intervention, as the program was designed to prevent rather than correct
oppositional behavior problems among the target children later in life. Only continued study of

these children will reveal whether the program was ultimately successful as a prophylactic

129

intervention, and whether later child outcomes are consistent with the variations in outcome noted
in the current work.
Summm and Future Directions

The present work is consistent with a small but growing body of literature on parent
training interventions that has emphasized the primacy of client investment and other treatment
process characteristics as predictors of treatment effectiveness. The work is a substantial
extension upon and reﬁnement of earlier research (Nye, et al. , in press) which identiﬁed maternal
treatment investment as a signiﬁcant predictor of child behavioral outcome among families
participating in the M.S.U. Multiple Risk Child Outreach Program. In the present study,
treatment investment, satisfaction, and expectation ratings of mothers and fathers who participated
in the program were examined as predictors of change in child behavior and authoritative
parenting. In addition, therapists’ level of satisfaction and expectations were examined, with the
expectation that these treatment process characteristics would be also be associated with treatment
outcome. Analyses showed that parent treatment investment, and parent and therapist satisfaction
and expectations, are related elements of the treatment process that are differentially associated
with change in child behavior and authoritative parenting reported at the end of the parent
training intervention. Further, the inﬂuence of these process characteristics does not appear to
be merely additive to the impact of pretreatment characteristics on outcome; under some
circumstances it mediates the inﬂuence of baseline parent functioning on subsequent treatment
success.

Implications for treatment. These ﬁndings indicate that variability in treatment outcome
is best accounted for by what goes on during treatment, and that such pretreatment characteristics
as parent psychopathology, family socioeconomic resources, level of child functioning, and

authoritative parenting style are of limited utility in predicting which families will experience

130

success as a result of parent training interventions. Although some of these baseline variables
may be potential disrupters of treatment, they generally do not in and of themselves predetermine
better or worse treatment outcomes. Instead, researchers and therapists alike would do well to
attend to aspects of the treatment process, in order to understand why some families are more
successful than others during parent training interventions. The current work provides evidence
that only parents who are observed to be more actively engaged during treatment and who follow-
through with homework assignments will show improvement in their parenting skills and, as a
result, will experience increased positive child behavior and decreased negative child behaviors.
These parents are also likely to report higher levels of satisfaction with the program, and to
expect that the intervention will continue to promote positive change. Similarly, therapists who
experience parents as being invested in treatment will also report high levels of satisfaction and
will anticipate that these families will continue to improve over time. Thus, parent investment
and parent and therapist satisfaction during the initial stages of therapy may serve as markers
regarding potential outcome, and intervention efforts must therefore include explicit attempts to
address deﬁcits in parent motivation early on. However, there is also evidence that continued
investment and motivation is necessary throughout treatment, particularly when more difﬁcult
personal and marital issues are addressed. This implies that successful interventions may require
continual monitoring of and ongoing attention to parent investment during the course of
treatment.

In applying these ﬁndings to a clinical setting, several approaches are suggested. First,
therapists must be actively involved in addressing not only targeted child behaviors, but also
parent attitudes and characteristics which may undermine treatment. Given the association
between pretreatment parent expectations and subsequent investment, initial sessions with parents

might need to explore and address possible reservations regarding the utility of the program, and

131

to then negotiate a therapeutic contract that gives the parents some sense of ownership regarding
the treatment. Further, to the extent that initial levels of positive child behavior may increase
parents’ investment in child-focused interventions, early work with families may need to focus
on ways to enhance the parent-child relationship and increase the frequency of positive exchanges
between parent and child. In addition, the disruptive impact of parent psychopathology to the
treatment process indicates that symptom reduction in the parents is a worthwhile goal at all
stages of an intervention. As noted earlier, parents may see the shift of focus from child
behavior problems to individual parent and marital difﬁculty as unnecessary and/or aversive.
Therefore, embedding this latter type of work within the larger fabric of the intervention may
help to reduce parents’ reactive resistance. Finally, evaluating parent satisfaction with and
reactions to the treatment process during the course of an intervention may assist therapists in
building on positive experiences and preventing more negative ones.

Further research is necessary to examine the potential interaction of other pretreatment
and treatment process characteristics in accounting for variations in parent training outcome.
Other relevant pretreatment characteristics include the initial age of the target child (Dishion &
Patterson, 1992), parent age (Nye, 1992) and negative life experiences (W ebster-Stratton, 1985;
Webster-Stratton & Hammond, 1990), as well as the level and quality of social support available
to parents (Dumas & Wahler, 1983). Additional potential pretreatment predictors of outcome
which would not have been relevant to the current sample but which have also been associated
with treatment success are single parent status (Webster-Stratton, 1985; Webster-Stratton &
Hammond, 1990), gender of the target child (Furey & Basili, 1988), and ethnic background of
the family (Holden, et al., 1990).

Regarding other treatment process predictors of outcome, therapist socioeconomic,

demographic, personality, and professional characteristics may be worthy of consideration.

132

Although there is conﬂicting data about the extent to which therapist characteristics inﬂuence
treatment success, there is some evidence that these may account for a modest but statistically
signiﬁcant percentage of the variance in treatment outcome (e.g. Crits-Christoph & Mintz, 1991;
Lambert, 1989: Nelson & Taylor, 1993). Patterson and his colleagues (Chamberlain, Patterson,
Reid, Kavanaugh, & Forgatch, 1984) found that level of therapist experience was associated with
aspects of the treatment process, as novice therapists encountered higher levels of resistance from
parents during sessions than did more experienced therapists. However, therapist characteristics
may have more of an impact on the process than on the outcome of therapy (e. g. Gomes-
Schwartz, 1978; Howard, Orlinsky, & Perilstein, 1976), and therapist experience alone may not
determine treatment effectiveness (e.g. Stein & Lambert, 1984). Orlinsky and Howard (1980)
also suggest that therapist characteristics are not systematically associated with outcome, but may
instead inﬂuence the range of clients with whom a therapist is effective. Similarly, Howard,
Orlinsky, and Perilstein (1976) have differentiated between "relational" and "instrumen "
aspects of therapy, and suggest that therapist personality characteristics are more associated with
the latter than the former. In this model, therapist characteristics are likely to inﬂuence clients’
affective experiences during therapy, but not to signiﬁcantly affect successful resolution of
identiﬁed problems. Other research also suggests that therapist characteristics, like parent
characteristics, may inﬂuence treatment outcome primarily via their impact on the treatment
process. Speciﬁcally, this research has found an association between therapist personality and
other characteristics and the quality of the therapeutic alliance, which in turn may affect clients’
eventual success in therapy (Frieswyk, Allen, Colson, Coyne, Gabbard, Horwitz, & Newsom,
1986; Lambert, Shapiro, & Bergin, 1986; Luborsky, McLellan, Woody, O’Brien, & Auerbach,
1985; Nelson & Taylor, 1992). Type of therapeutic alliance has also been associated with

whether clients complete or prematurely terminate treatment (Tryon & Kane, 1993).

133

Further research is also necessary to evaluate the interplay between pretreatment and
treatment process characteristics over the course of parent training interventions. In examining
the inﬂuence of parent treatment investment and parent and therapist satisfaction and expectation
ratings on outcome, the present work used global, composite scores for these process constructs.
However, investment, satisfaction, and expectations are likely to be multi-faceted and multiply-
determined characteristics of the treatment process. In the present report, parent treatment
investment scores were comprised of ratings on the parents’ completion of homework
assignments, and well as therapist observations of and reactions to parent behaviors within
sessions. These latter ratings included the extent to which participating parents talked during the
session, and were withdrawn or outgoing, open to new ideas, hostile or friendly toward the
therapist, and resistant or facilitative within each session. In addition, the therapist reported on
how she/he felt toward each parent, and how well s/he worked with each parent during the
sessions. Although these various indices were determined to be statistically related (and were
therefore combined into one composite investment score), it is possible that certain of these
treatment investment indices were differentially associated with measures of pretreatment parent
and family functioning and with change in child behavior and parenting. Other studies have
suggested that homework compliance is a separate dimension of treatment investment (Patterson
& Chamberlain, 1988) that is not necessarily related to within-session process characteristics
(Nelson & Borkovec, 1989). There is also evidence that various aspects of parent engagement
during sessions may be differentially associated with pretreatment parent and family functioning,
with other aspects of the treatment process, and with treatment outcome. Patterson and
Chamberlain (in press) differentiated between "I can’t" and "I won’t" statements from parents in
response to therapists’ interventions, and found that these statements represented two qualitatively

different types of resistance associated with different pretreatment parent and family

134
characteristics. In this regard, "I can’t" statements were associated with baseline depression and
antisociality for both mothers and fathers, and father "I won’t" statements were also related to
paternal antisociality. Further, these two types of resistance were differentially predictive of
therapists’ feelings about the parents, with the number of sessions required before termination,
and with the amount of session time spend discussing social learning topics. In interpreting these
ﬁndings, the authors suggest that:

the therapist’s ability to persist in teaching key skills in the face of the expressed

helplessness of the parent and the therapist’s own increasing discouragement or

dislike appears to account for much of the variance in the success of treatment

(Patterson, Dishion, & Chamberlain, 1993, p. 67).

Interestingly, these researchers have also observed that within-session resistance ("I won’t") is
not necessarily detrimental to treatment, but may actually be an indicator that parents are actively
engaged in the program (Patterson & Chamberlain, 1988). These authors propose that a
moderate level of struggle between parents and therapists is a necessary precursor to change in
parenting practices, but that resistance that is too extreme will result in less improvement in
parenting and more poor outcomes.

The use of composite scores to represent parents’ and therapists’ global satisfaction in the
present work also precluded an examination of whether there may be multiple components of
treatment satisfaction that are differentially associated with pretreatment, treatment process, and
treatment outcome characteristics. Some parents may have been quite satisﬁed with the child-
focused aspects of the intervention but relatively dissatisﬁed with the marital problem-solving
work, which might then have different implications for reported outcome. Parents’ and
therapists’ reactions to and feelings about one another might also be separate process dimensions
that uniquely contribute to both treatment investment and treatment outcome. Clearly, this would

be consistent with the research linking therapeutic alliance with outcome (Frieswyk, et al. , 1986;

Lambert, et al., 1986; Luborsky, et al., 1985; Nelson & Taylor, 1992). Further, these various

135

aspects of treatment satisfaction may be associated with different pretreatment parent and family
characteristics. In addition, the Patterson work demonstrates that direct observations of client and
therapist behavior in treatment are important to understanding the treatment process
(Chamberlain, et al., 1984).

There is also evidence that mother and father treatment process characteristics may have
different implications for treatment outcome (e.g. Patterson & Chamberlain, 1994, and should
therefore be examined separately. Such comparisons were not conducted in the present work,
because of the protocol’s restriction on the number of fathers who participated in treatment.
However, it is interesting to note that father psychopathology was signiﬁcantly correlated with
parent investment during the second half of the treatment, such that families in which the fathers
were more disturbed were less invested in the marital problem-solving work. Patterson and
Stoolmiller (in press) report that changes in both maternal and paternal resistance inﬂuence
changes in mothers’ disciplinary strategies, indicating that fathers’ behavior in treatment
contributed to mothers’ improvement in parenting. Patterson and Chamberlain (1992) observed
that maternal psychopathology and socioeconomic disadvantage accounted for mothers’ initial
resistance in treatment, but were not related to maternal investment at the end of the intervention.
In contrast, paternal psychopathology and disadvantage impaired fathers’ investment throughout
treatment.

These latter ﬁndings also indicate that variability in treatment process characteristics over
time should be examined more extensively. Within the current treatment sample, overall parent
investment was observed to drop from mid-treatment to termination. It was hypothesized that
this decrease in investment was attributable to changes in the intervention protocol, whereby
contacts with the parents were less frequent but potentially more confrontational and emotionally-

charged. However, there was also some evidence that higher parent investment across the span

136

of the intervention had particular implications for child outcome, whereas early parent investment
did not. Therefore, changes in treatment investment during the course of an intervention may
be an important determinant of treatment success, and examining treatment process characteristics
on a microcosmic, session-by-session level may be necessary in order to more clearly delineate
how these are related to one another and to treatment outcomes over time. Patterson &
Chamberlain (1994) have advanced a "struggle hypothesis", which proposes that successful
interventions should be marked by an observable increase in parent resistance during the middle
of treatment, followed by a signiﬁcant reduction by termination once the resistance has been
worked through (Patterson & Chamberlain, 1992). This has been supported by research
indicating that parents who are successful in parent training exhibit a progressive increase in
resistance from beginning to mid-treatment sessions, and a subsequent tapering off of resistance
to termination (Stoolmiller, Duncan, Bank, & Patterson, 1993). Interestingly, this model seems
to parallels Tracey (1989) identiﬁcation of a curvilinear pattern in satisfaction levels among
clients experiencing successful outcomes, whereby clients report less satisfaction during the
middle of treatment as a result of dealing with more difﬁcult issues, but then express increased
satisfaction when these issues have been resolved by termination. The study of investment and
satisfaction at two, rather than three, time points during treatment in the present work prevented
this type of analysis.

Rosen and Proctor (1981) have suggested that treatment outcome is multifaceted, and is
intimately imbedded within the treatment itself. They differentiate ultimate from instrumental and
intermediate outcomes, and argue that each of these is an important characteristic of treatment.
Ultimate outcomes are those which are the speciﬁc identiﬁed goals of an intervention. In
contrast, instrumental outcomes are thought to contribute to other outcomes without additional

intervention. In the present study, change in authoritative parenting style may be considered the

137

instrumental outcome which was expected to facilitate the attainment of the ultimate outcome --
the reduction and prevention of antisocial child behavior. Intermediate outcomes are client
characteristics which are considered to be necessary for interventions to continue or to be
successfully employed. In this case, levels of parent treatment investment, satisfaction, and
psychopathology may all have been intermediate outcomes which contributed to the eventual
attainment of initial changes in parenting and subsequent improvement in child behavior. Thus,
this model proposes that outcomes occur throughout treatment. Therefore, process and outcome
research must go beyond assessing client outcomes at termination and follow-up, and should
recognize the "continuum of client change" that occurs during an intervention. In the present
study, changes in child behavior, authoritative parenting, parent investment, and parent and
therapist satisfaction and expectations were evaluated at mid-treatment, as well as at termination
and, in some cases, follow-up. However, this work represents only one small step toward
understanding the relationships among treatment processes and outcomes.

In expanding our knowledge of the multiple inﬂuences on treatment outcome in parent
training interventions, one additional path merits exploration. In this and other studies,
pretreatment parent and family characteristics have tended to be examined primarily as static
states that may predispose families to better or worse outcomes at termination. But it is possible,
if not highly likely, that some of these characteristics are also modiﬁed during the course of
treatment and should therefore be considered not only as other indices of treatment outcome, but
as more dynamic contributors to treatment success. This is supported by research on parent
training programs with a relationship enhancement or marital problem-solving component. These
studies have shown that parent training programs that address problems in both parenting
practices and the marital relationship are more successful in reducing deviant child behavior than

those which focus more exclusively on modifying parent disciplinary practices (e. g. Dadds, et

138

al., 1987; Sayger, Horne, & Glaser, 1993), suggesting that changes in marital satisfaction or
adjustment may coincide with improved child outcome. Thus, marital conﬂict may not merely
be a pretreatment disrupter variable, but also a more dynamic inﬂuence on child outcome over
the course of an intervention. Similarly, the incorporation of "parent enhancement therapy"
focusing not only on parenting behaviors but also addressing problems with parents’ personal and
marital adjustment, extra-familial relationships, and perceptions of the target child’s behavior, has
proven more effective than a more traditional parent training approach in changing child behavior
and parenting practices (Griest, Forehand, Rogers, Breiner, Furey, & Williams, 1982). Thus,
parent training programs that explicitly target parent individual and marital distress appear to
promote more marked improvement in child behavior, suggesting that changes in these
pretreatment parent characteristics during the course of the intervention may account for some
of the variance in child outcome. Further, even more exclusively child-focused intervention
protocols that do not systematically address problems in parent and marital functioning may
nonetheless yield non-speciﬁc treatment effects, such that problems in parents’ personal and/or
marital adjustment are also alleviated during the course of treatment (e. g. Scovern, Bukstel,
Kilman, Laval, Busemeyer, & Smith, 1980; Winans & Cooker, 1984). Therefore, in order to
more adequately capture the interplay among parent, child, family, and treatment process
characteristics in predicting treatment success, future research should examine whether changes
in such pretreatment correlates of problem child behavior as parent distress and marital
dissatisfaction occur during parent training interventions, and how these changes may be

associated with treatment investment, satisfaction, and outcome.

APPENDICES

APPENDIX A

PEARSON CORRELATIONS BETWEEN PARENT PSYCHOPATHOLOGY VARIABLES

AND PSYCHOPATHOLOGY CLUSTERS

139
APPENDIX A

Pearson Correlations Between Parent Psychopathology Variables and Psychopathology Clusters

 

 

Parent Psychopathology Variables Cluster
Father LAPS ASB BECK HAMW FPROBS
LAPS .54** .35" .39** .80**
ASB .25* .22* .71**
BECK .31** .68**
HAMW .67* :1:

Mother
LAPS ASB BECK HAMW MPROBS
LAPS .55** .30** .29** .75**
ASB .37** .29** .78**
BECK .24* .68M
HAMW .64**

Note. LAPS=Lifetime Alcohol Problems; ASB=Antisocial Behavior; BECK=Beck Current

Depression; HAMW=Hamilton Worst-Ever Depression; FPROBS=Father Psychopathology

cluster; MPROBS =Mother Psychopathology cluster

*p< .05. **p< .01.

APPENDIX B

PEARSON CORRELATIONS AMONG FAMILY DEMOGRAPHIC CHARACTERISTICS

AND THE FAMILY RESOURCES CLUSTER

140
APPENDIX B

Pearson Correlations Among Family Demographic Characteristics and the Family Resources

 

QM
Demographic Variables Cluster
MOMED DADED INCOME SES RESOURCES
MOMED .46** .39** .60** .78”
DADED .38** .59** .77**
INCOME .51** .73**
SES . 86* *

Note. MOMED=Mothers’ Education; DADED=Fathers’ Education;
RESOURCES=Family Resources cluster (Income, SES, and Education).

*p< .05, two-tailed. **p< .01, two-tailed.

APPENDIX C

FAMILY DEMOGRAPHIC AND PARENT PSYCHOPATHOLOGY CHARACTERISTICS

OF THE TREATMENT SAMPLE

141
APPENDIX C

Family Demographic and Parent Psychopathology Characteristics of Treatment Sample (N =29)

 

Characteristic M £2

 

Family Demographics

Income $26,965 17,750
Socioeconomic Status 288 122
Mother Age (yrs) 29.2 4.6
Father Age (yrs) 32.1 5.6
Mother Education (yrs) 12.8 1.6
Father Education (yrs) 12.4 1.6

Mother Psychopathology

Lifetime Alcohol Problems 9.6 1.7
Lifetime Antisocial Behavior 14.7 9.4
Current Depression (Beck) 4.7 3.8
Worst Ever Depression (Hamilton) 15.2 8.4
Father Psychopathology
Lifetime Alcohol Problems 10.6 1.8
Lifetime Antisocial Behavior 24.1 14.3
Current Depression (Beck) 3.1 2.8
Worst Ever Depression (Hamilton) 14.2 8.3

 

Note. Socioeconomic status was measured by the Duncan TSE12 index (Stevens & Featherman,

1981).

APPENDIX D

FACTOR LOADINGS OF OBSERVER IMPRESSIONS INVENTORY ITEMS

IN THE AUTHORITATIVE PARENTING SCALE

BASED ON PRINCIPAL COMPONENTS ANALYSIS

142
APPENDIX D

Factor Loadings of Observer Impressions Inventory Item_s in the Authoritative Parenting Scale

 

Based on Principal Components Factor Analysis

 

Observer Impressions Item (item #) Factor Loading
Expects child to behave nicely toward others (40a) .61
Expects child to use good manners (40b) .60
Is a positive and reinforcing parent (40c) .77
Encourages/sees humor in aggressive/antisocial behavior (40d) -.52
Is overly permissive, laissez-faire, negligent (41b) -.63
Is erratic, inconsistent, haphazard (41c) -.79
Is consistent, even-handed, ﬁrm when necessary (41d) .86
Seems in good control of target child (41c) .78
Seems to have little/no control of behavior of target child (41f) -.77
Uses nagging or nattering to get compliance (41h) -.52
Expresses anger/hostility while disciplining (41) -.65
Friendly relations between parent and child (42a) .77
Parent seemed distant/detached from child (42d) -.59

Parent treats target child with respect (42k) .71

APPENDIX E

PEARSON CORRELATIONS BETWEEN SELECTED REVISED BELSKY RATINGS

OF PARENTING BEHAVIOR AND AUTHORITATIVE PARENTING ITEMS

ON THE OBSERVER IMPRESSIONS INVENTORY

FOR THE LARGER TREATMENT SAMPLE

38-28 .5. v9... 35-08 .8. Va...

 

 

mm. a .- 1.3.. ...mm. 2 .- was».
1.8.- E. :.- 2.- 3%. consamomaom ”sesame
...mm.. 2. .50. 8.- *8. 838$ 338m
to- ...X. :3..- 8. :3. as? ”>38“.

Ev ass 6:. sec 0:. 835 3:. 805 69. 82c scenic ”seen; 328
Em: “Sim—EU 23am channommmj w:_o~o.«:mom\u>§mom

 

allJlmvH: o_mEa_.m “cos—Smock 53mg 05 .28 39.52: 8232 E— 83030 2: no 25: 528$

88— beagﬁ 823895 SPEED

 

 

 

m X—Dmen?‘

m3

APPENDIX F

PEARSON CORRELATIONS BETWEEN PARENT AND THERAPIST SATISFACTION

ITEMS AND PARENT AND THERAPIST EXPECTATIONS

144
APPENDIX F

Table 14

Pearson Correlations Between Parent Satisfaction Items and Parent Expectations at Post-tests

 

Parent Satisfaction Item Parent Expectations Rating
Target child’s behavior has improved ,45****
More positive feelings about target child ,35****
Therapist understands problems raised in sessions ,40****
Positive effect on parents’ relationship ,35****
Positive effect on parent’s relationship w/ target child .48****
Positive effect on parent personally 40”“
Parent liked the therapist ,33****
Satisﬁed overall with the treatment .42****

**p<.01, two-tailed. ***p<.001, two-tailed. ****p<.0001, two-tailed.

145
APPENDIX F

Table 15

Pearson Correlations Between Therapist Satisfaction Items and Therapist Expectations at Post-tests

 

Therapist Satisfaction Item Therapist Expectations Rating
Target child’s behavior has improved .55****
More positive feelings about target child .43**
Parents understood treatment directions/techniques .50****
Positive effect on parents’ relationship .44**
Positive effect on mother’s relationship w/ target child .69****
Positive effect on father’s relationship w/ target child .42**
Positive effect on mother personally .43**
Positive effect on father personally .53***
Therapist liked the mother .38**
Therapist liked the father .27
Satisﬁed overall with the treatment .68****

**p< .01, two-tailed. ***p< .001, two-tailed. ****p< .0001, two-tailed.

APPENDIX G

PEARSON CORRELATIONS AMONG MOTHER, FATHER, AND THERAPIST RATINGS

OF CHILD BEHAVIOR, AND THE COMPOSITE CHILD BEHAVIOR SCORES

147

Appendix G
Table 17
Pearson Correlations Among Mother, Father, and Therapist Ratings of Negative Child Behavior.

and the Composite Negative Child Behavior Score

 

 

 

Child Behavior Rating
Negative Behavior Measure Mother Father Therapist Composite
Pretreatment (Baseline)
Mother -- .53** -.03 .59***
Father -- .07 .69***
Therapist -- .70***
Post-test l (Mid-treatment)
Mother -- .25 .37"‘ .70***
Father -- .27 .72***
Therapist -- .77***
Post-test 2 (Termination)
Mother -- .34* .29 .79***
Father -- . 16 .69***
Therapist - .65***
Post-test 3 (Follow-up)
Mother -- . 10 -- .77***
Father -- -- ’72:“

 

*p< .05, one-tailed. **p< .01, one-tailed. ***p< .001, one-tailed.

APPENDIX H

TESTS OF PARENT TREATMENT INVESTMENT AS A MODERATOR BETWEEN

PRETREATMENT AND POST-TEST CHILD BEHAVIOR AND PARENTING

148

Subsequent to regression analyses testing the contribution of parent treatment investment
in predicting child behavior and parenting at the various post-test intervals, analyses were also
conducted to examine whether investment moderated the relationship between pretreatment and
post-test child behavior and parenting. This decision was based on earlier analyses with this
treatment sample (Nye, et al., in press), which indicated that signiﬁcant child behavior change
occurred only among families in which the mothers were more highly invested in the
intervention. In contrast, child behavior at post-test was observed to be essentially unchanged
from pretreatment level among families with less-invested mothers. Similar trends were expected
in the current work, with parent treatment investment anticipated to moderate the inﬂuence of
pretreatment child behavior and authoritative parenting on subsequent child behavior and
parenting, respectively. In particular, child and parenting outcomes were expected to be
relatively independent of pretreatment child and parent functioning among highly invested
parents, whereas the relationships between pretreatment and post-test levels of child behavior and
parenting were expected to be stronger when parent investment was lower. Because of the small
sample size, a more lenient 10% alpha level was used in initial analyses testing for an interaction
between parent investment and pretreatment child or parent functioning.

8%

In order to test parent treatment investment as a moderator variable, interaction scores
were generated for the product of parent investment and the individual pretreatment child and
parenting means. The relevant interaction term was then entered at Step 3 of the regression
equations, after pretreatment child behavior or parenting (Step 1) and early or cumulative parent
investment (Step 2). These analyses identiﬁed Cumulative Parent Investment as a moderator
variable for Positive Child Behavior and Authoritative Parenting at termination, and Negative

Child Behavior at the six-month follow-up. Early Treatment Investment was not observed to

149

moderate any of the mid-treatment outcomes.

Regarding Positive Child Behavior at termination, an interaction between Cumulative
Investment and pretreatment Positive Behavior was identiﬁed (A32: .08, p= .09) (Table 18).
The nature of this interaction was examined by dividing treatment families into two groups, based
on a median split on parents’ Cumulative Treatment Investment scores. Separate regression
analyses were then run for these two groups, examining pretreatment Positive Child Behavior as
a predictor of Positive Child Behavior at termination. These analyses indicated a signiﬁcant
relationship between pretreatment and termination Positive Child Behavior for families who
showed less investment in the intervention protocol (32:53, B(1,12)=13.56, p< .01), while
level of Positive Child Behavior at pretreatment was not predictive of Positive Child Behavior
at termination for families who demonstrated a higher level of treatment investment (Bz=.02,
B(1,13)=.22, p=n.s.). As shown in Figure 2, Positive Child Behavior at termination was
consistently high among more invested families, regardless of initial level of positive behavior.
In contrast, termination level of Positive Child Behavior for less invested families appears to have
been linked with pretreatment positive behavior, with low initial positive behavior associated with
lower positive behavior at termination. Thus, Cumulative Parent Treatment Investment
moderated the impact of pretreatment Positive Child Behavior on Positive Child Behavior at
termination for families in which parents demonstrated a higher level of investment throughout
treatment, indicating that only more highly invested families experienced signiﬁcant and
consistent change in positive child behavior regardless of pretreatment level.

A signiﬁcant interaction between pretreatment Negative Child behavior and Cumulative
Parent Treatment Investment was also found for the prediction of Negative Behavior at follow-up
(A3}: .09, p < .05) (Table 19). Therefore, treatment families were again divided into two groups

based on a median split of the Cumulative Investment score, and the relationship of pretreatment

150
Table 18

Hierarchical Regression Analyses of Parent Treatment Investment as a Moderator of Pretreatment

Positive Child Behavior on Positive Child Behavior at Termination

 

Predictor Variable B S_ B B _R_2
Step One .15**
Baseline Positive Child Behavior .41 .19 .39**
Step Two .31***
Baseline Positive Child Behavior .21 .19 .20
Cumulative Treatment Investment .37 .15 .45**
Step Three .39***
Baseline Positive Child Behavior .26 .18 .24
Cumulative Treatment Investment .32 .15 .38**
Baseline Behavior X Investment -. 12 .07 -.28*

 

*p< .10. **p< .05. ***p< .01.

151

T2 Positive Behavior

555555 = Less investment

.70

m = More investment
.60
.50
.40
.30
.20
.10
.00

-.10

-.20

-.30

 

-.40

-.50

-.60

-.7O

-.80

 

-.90

-1.00

__L_L_L_L_L_L_L_L_L_L_L_L_J__L_L_L._L_L_L_l__l__l__l__l__l__L_l__L_L_l__L_l__l__L.J__l__L_|

- 1 0 + 1
Pretreatment Positive Behavior
Figure 2. Relationship between Pretreatment and Termination (T2) Positive Child Behavior based

on Cumulative Parent Treatment Investment.

152
Table 19

Hierarchical Regression Analyses of Parent Treatment Investment as a Moderator of Pretreatment

Negative Child Behavior on Negative Child Behavior at Follow-up

 

Predictor Variable g _SE B. B _R_2
Step One _25***
Baseline Negative Child Behavior .60 .20 .50***
Step Two .39***
Baseline Negative Child Behavior .56 .18 .47***
Cumulative Treatment Investment -.34 .14 -.37**
Step Three 39*"
Baseline Negative Child Behavior .65 .18 .55***
Cumulative Treatment Investment -.37 .14 -.40**
Baseline Behavior X Investment -.18 .09 -.32**

 

*p<.10. **p<.05. ***p<.01. ****p<.001.

153

to follow-up Negative Child Behavior was examined separately for each group. These regression
analyses identiﬁed pretreatment Negative Child Behavior as a signiﬁcant predictor of subsequent
Negative Behavior at follow-up for the group of families showing lower cumulative treatment
investment (32: .37, B(l ,l2)=7.08, p < .05), but not for those families where parent investment
was higher across the span of the intervention (32: .22, B(1,13)=3.79, p=n.s.). The graph of
these relationships (Figure 3) demonstrates a consistently lower level of negative behavior among
more highly invested families, regardless of the target child’s pretreatment behavior problems.
In contrast, level of negative child behavior at follow-up among target children in less invested
families was strongly associated with initial negative behavior, with more troublesome
pretreatment behavior predicting more problematic behavior six months post-termination. Thus,
families demonstrating greater investment in the intervention protocol were consistently able to
effect improvement in negative child behavior by follow-up, independent of the initial level of
negative behavior observed in the target child. For less invested families, child behavioral
outcome was dependent upon pretreatment level.

Cumulative Treatment Investment was also observed to moderate the relationship between
pretreatment and termination levels of Authoritative Parenting (A32: .04, p= .07) (Table 20).
When the relationship between Parenting at pretreatment and termination was examined separately
for families showing differential levels of treatment investment, pretreatment Parenting was found
to be predictive of Parenting at termination for less invested families (32: .29, B(1,12)=4.82,
p< .05), but not for families who exhibited a greater level of investment throughout the
intervention (B’=.05, B(1,13)=.72, p=n.s.). As depicted in Figure 4, level of Authoritative
Parenting at termination for highly invested families was relatively independent of pretreatment
parenting, as parenting outcome for these families was consistently high regardless of initial

parent functioning. Among less invested parents, authoritative parenting skills observed at the

154
Table 20

Hierarchical Regression Analyses of Parent Treatment Investment as a ModeLator of Pretreatment

 

Authoritative Parenting on Authoritative Parenting at Termination

 

Predictor Variable _B §1_3 13. E E
Step One ,22***
Baseline Parenting .60 .21 47*“
Step Two .65*****
Baseline Parenting , 35 .15 23“
Cumulative Treatment Investment .81 .14 .6‘8*****
Step Three gonna:
Baseline Parenting .36 .14 “29*“
Cumulative Treatment Investment .77 .14 .65*****
Baseline Parenting X Investment -. 14 .07 -.21*

 

*p_<.10. **p<.05. ***p<.01. ****p<.001. *****p<.0001.

155

T3 Negative Behavior

555555 =Less investment 3.“:
1.20
m =More investment
1.10
1.00
.90
.80
.70
.60
.50
.40
.30
.20
.10
.00
-.10
-.20
-.30
-.40

-.50

-.60

_...l_.l__L_l__L_l__L_.|__L..L._.|.__L._l.__l__.l_.1...l__l__l__l__L_L_L_L_L_l__L..l__L_L..|._._L_l.__l.__l__L_L_L_I

 

 

- 1 0 + 1
Pretreatment Negative Child Behavior
Figpre 3. Relationship between Pretreatment and Follow-up (T3) Negative Child Behavior based

on Cumulative Parent Treatment Investment.

156

T2 Parenting F
555555 =Less investment

  
 
 

.70 ”=More investment
.60
.50
.40
.30
.20
.10
.00

-. 10

-.20

-.30

-.40

-.50

-.60

—.70

-.80

-.90

-1.00

-1.10

 

__L_L_.|__L_l__l__L_l_.J__L_L.L.J__.l__l__L_L_L_L_L_1__L_L_L_L_L_L.J__.1__L_L_L_L_1__1__L_L_L_L.

 

l I l
-l 0 +1

Pretreatment Authoritative Parenting
Figpre 4. Relationship between Pretreatment and Termination (T2) Authoritative Parenting based

on Cumulative Parent Treatment Investment.

157

end of the intervention were strongly associated with initial parenting, indicating relatively less
change within this subset of treatment families. Level of authoritative parenting changed
consistently and signiﬁcantly from pretreatment to termination only among more highly invested
families.
Discussion

Earlier regression analyses identiﬁed parent treatment investment as a signiﬁcant predictor
of improvement in positive and negative child behaviors, and authoritative parenting. The
interaction effects described herein indicate additional relationships that are important to an
understanding of observed variability in outcome among this treatment sample. Although all of
the families in the sample completed the entire intervention protocol, differential treatment
effectiveness was observed relative to parent treatment investment. In this regard, Cumulative
Treatment Investment was identiﬁed as a moderator between pretreatment and termination levels
of positive child behavior and authoritative parenting, and between negative child behavior at
pretreatment and follow-up. Thus, parents’ overall investment across the span of the intervention
inﬂuenced the extent to which child and parenting behavior changed from initial pretreatment
levels. For families in which cumulative investment was high, child and parenting outcomes were
observed to be unrelated to pretreatment levels, suggesting a substantial treatment effect within
this group. In contrast, pretreatment and termination levels of child behavior and parenting style
were highly correlated within the less invested group, suggesting negligible improvement in
targeted parent and child behaviors from baseline levels among families whose overall investment
was lower. Thus, cumulative treatment investment was differentially associated with better or
worse outcomes in child behavior and parenting among the families in this sample, with
signiﬁcant improvement found primarily among the more highly invested families.

Although these interactions must be interpreted with caution because of the more liberal

158

effect size criterion used to identify them, their potential signiﬁcance should not be overlooked.
The fact that cumulative, but not early, level of parent investment acted as a moderator of
pretreatment functioning suggests that level of investment in the second half of treatment may
have had particular signiﬁcance for eventual outcome at termination. Again, the second half of
treatment addressed issues that were typically more intense and difﬁcult for the parents, and
treatment investment was observed to decrease overall during this phase of the intervention.
What these ﬁndings involving cumulative parent investment suggest is that parents who were
relatively more able to maintain a higher level of investment during this phase of treatment (and
therefore had higher cumulative investment scores) were more successful at effecting changes in
positive and negative child behavior and in authoritative parenting independently of baseline levels
of child and parent functioning. Although the precise reason for this ﬁnding cannot be
determined, it may indicate that these parents were better able to process and resolve personal
and interpersonal (marital) issues which might otherwise have interfered with treatment progress.
These families, by virtue of their greater willingness to address difﬁcult personal and marital
issues, subsequently experienced more positive outcomes at termination. This hypothesis is
consistent with research indicating that marital conﬂict impedes outcome in parent training
programs unless successfully addressed as part of the intervention, and that child outcome is
enhanced in parent training programs that attend to personal and marital difﬁculty (Dadds, et al.,
1987a; Griest, Forehand, Rogers, Breiner, Furey, & Williams, 1982; Sayger, Horne, & Glaser,

1993). Again, however, this explanation must be considered to be quite tentative at this time.

LIST OF REFERENCES

LIST OF REFERENCES

Adler, R., & Raphael, B. (1983). Children of alcoholics. Australian and New Zealand
Journal of Psychiatry, 12, 3-8.

Alexander, J.F., & Parsons, B.V. (1973). Short-term behavioral intervention with

delinquent families: Impact on family process and recidivism. Journal of Abnormal Psychology,
l_8, 219-225.

Alterman, A.I., Bridges, R.K., & Tarter, RE. (1986). Drinking behavior of high risk
college men: Contradictory preliminary ﬁndings. Alcoholism, 19, 305-310.

Alterman, A.I., & Tarter, RE. (1983). The transmission of psychological vulnerability:
Implications for alcoholism etiology. Journal of Nervous an_d Men_tal Disea_se, m, 147-154.

 

Alvy, K.T. (1991). Parent training as a prevention strategy. In D. Robbins (Ed.),

Parent training is prevention: Preventing alcohol and other drug problemsamong youth in the
family. Rockville, MD: US. DHHS.

Anderson, S. A., Atilano, R. B., Bergen, L. P., Russell, C. S., & Jurich, A. P. (1985).
Dropping out of marriage and family therapy: Intervention strategies and spouses’ perceptions.
The American Journal of Family Therapy, _1_3, 39-54.

Baekeland, F., & Lundwall, L. (1975). Dropping out of treatment: A critical review.
Psychological Bulletin, _8_2, 738- 783.

Bandura, A. (1977). Self-efﬁcacy: Toward a unifying theory of behavior. Psychological
Review, &, 191-215.

Bandura, A. (1989). Human agency in social cognitive theory. American Psychologist,
_4, 1175-1184.

Baron, R. M., & Kenny, D. A. (1986). The moderator-mediator distinction in social
psychological research: Conceptual, strategic, and statistical considerations. Journal of

Personality and Social Psychology, 51, 1173-1182.

Beck, A. T., & Beck, R. W. (1972). Screening depressed patients in family practice:
A rapid technique. Postgraduate Medicine, §_2_, 81-85.

Belsky, J. (1981). Early human experience: A family perspective. Developmental
Psychology, ﬂ, 3-23.

Belsky, J. (1984). The determinants of parenting: A process model. Child
Development, 3, 83-96.

Belsky, J. , Robins, E. , & Gamble, W. (1984). The determinants of parental competence.
In M. Lewis (Ed.), Beyond the dyad. NY: Plenum Press.

159

160

Bennett, L. A., & Wolin, S. J. (1987). Couples at risk for transmission of alcoholism:
Protective inﬂuences. family Process, 111-129.

Billings, A. G., & Moos, R. H. (1983). Comparisons of children of depressed and
nondepressed parents: A social-environmental perspective. Journal of Abnormal Psychology,
_1_1, 463-486.

Bloch, 8., Bond, G., Qualls, B., Yalom, I., & Zimmerman, E. (1976). Patients’
expectations of therapeutic improvement and their outcomes. American Journal of Psychiatgz,
_1_33_, 653-657.

Block, J. H., Block, J ., & Gjerde, P. F. (1986). The personality of children prior to
divorce: A prospective study. Child Development, _51, 827-840.

Block, J. H., Block, J., & Morrison, A. (1981). Parental agreement-disagreement on
child-rearing orientations and gender related personality correlates in children. Child
Development, ﬂ, 965-974.

Bowden, C. L., Schoenfeld, L. S., & Adams, R. L. (1980). Mental health attitudes and
treatment expectations as treatment variables. Journal of Clinical Psychology, 315, 653-657.

Brody, G. H., & Forehand, R. (1986). Maternal perceptions of child maladjustment as
a function of the combined inﬂuence of child behavior and maternal depression. Journal of

Consulting and Clinical Psychology, 55, 237-240.

Brody, G. H., Pillegrini, A. D., & Sigel, I. E. (1986). Marital quality and mother—child
and father-child interactions with school-aged children. Developmental Psychology, B2,
291-296.

Brown, S. A., Creamer, V. A. & Stetson, B. A. (1983). Adolescent alcohol expectancies
in relation to personality traits and adolescent drug use. Journal of Abnormal Psychology, 9B,
117-121.

Burling, T. A., Reilly, P. M., Moltzen, J. O., & Ziff, D. A. (1989). Self-efﬁcacy and
relapse among inpatient drug and alcohol abusers: A predictor of outcome. Journal of Studies
on Alcohol, _S_Q, 354-360.

Burns, D. D., & Nolen-Hoeksema, S. (1991). Coping styles, homework compliance,
and the effectiveness of cognitive-behavioral therapy. Journal of Consulting and Clinical
Psychology, _5_9, 305-307.

Cernkovich, S. A., & Giordano, P. C. (1987). Family relationships and delinquency.
Criminology, 2_5, 295-321.

Chamberlain, P., Patterson, G. R., Reid, J. B., Kavanaugh, K., & Forgatch, M. S.
(1984). Observation of client resistance. Behavior Therapy, Q, 144-155.

Chamberlain, P., & Ray, J. (1988). The therapy process code: A multidimensional

161

system for observing therapist and client interactions in family treatment. In R.J. Prinz (Ed.),
Advances in behyioral assessment of children and families (Vol. 4, pp. 189-217). Greenwich,
CT: JAI Press.

Chassin, L., Pillow, D. R., Curran, P. J., Molina, B. S. G., & Barrera, M. (1993).
Relation of parental alcoholism to early adolescent substance use: A test of three
mediating mechanisms. Journal of Abnormal Psychology, 1_02, 3-19.

Chassin, L., Rogosch, F., & Barrera, M. (1991). Substance use and symptomatology
among adolescent children of alcoholics. Journal of Abnormal Psychology, 10B, 449-463.

Christensen, A., Phillips, 8., Glasgow, R, E. & Johnson, S. M. (1983). Parental
characteristics and interactional dysfunction in families with child behavior problems.
Journal of Abnormal Child Psychology, 11, 153-166.

Christiansen, B. A., & Goldman, M. S. (1983). Alcohol-related expectancies in relation
to personal and parental drinking patterns. Journal of Consulting and Clinical Psychology, 5_1,
249-257.

Clark, D. B., & Baker, B. L. (1983). Predicting outcome in parent training. Journal
of Consulting and Clinical Psychology, 5_l, 309-311.

Clark, D. B., Baker, B. L., & Heifetz, L. J. (1982). Behavioral training for parents of
mentally retarded children: Prediction of outcome. American Journal of Mental Deﬁciency, 8_7,
14-19.

Cloninger, R. (1987). Neurogenetic adaptive mechanisms in alcoholism. Science, 2_6,
410-416.

 

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

Cohen, J. (1987). Statistical mwer analysis for the behayioral scienpga. Hillsdale, NJ:
Lawrence Erlbaum Associates.

Collins, J. F., & Hyer, L. (1986). Treatment expectancy among psychiatric inpatients.
Journal of Clinical Psychology, 4;, 562-569.

Collins, J. J. (1981). Alcohol careers and criminal careers. In J.J. Collins (Ed.),
Drinking and crime. New York: Guilford Press.

Cotton, N. S. (1979). The familial incidence of alcoholism: A review. Journal of
Studies on Alcohol, 40, 89-116.

Crits-Christoph, P., & Mintz, J. (1991). Implications of therapist effects for the design
and analysis of comparative studies of psychotherapies. Journal of Consulting and Clinical
Psychology, 3, 20-26.

Cummings, E. M., Iannotti, R. J., & Zahn-Walker, C. (1985). Parental characteristics

162

and interactional dysfunction in families with child behavior problems: A preliminary
investigation. Journal of Abnormal Child Psychology, _1_1 , 153-166.

Cummings, E. M., Zahn-Walker, C., Radke-Yarrow, M. (1981). Young children’s
responses to expressions of anger and affection by others in the family. Child Development, ﬂ,
1274-1282.

Cummings, E. M., Zahn-Walker, C., & Radke-Yarrow, M. (1984). Developmental
changes in children’s reactions to anger in the home. Journal of Child Psychology and Psychiatg,
2_5, 63-73.

Dadds, M. R. (1987). Families and the origins of child behavior. Family Process, _2_6,
341-357.

Dadds, M. R., Sanders, M. R., Behrens, B. C., & James, J. E. (1987). Marital discord
and child behavior problems: A description of family interactions during treatment. Journal

of Clinical Child Psychology, 1a, 192-203.

 

Dadds, M. R., Schwartz, S. & Sanders, M. R. (1987). Marital discord and treatment
outcome in behavioral treatment of child behavior problems. Journal of Consulting and Clinical
Psychology, ﬂ, 396-403.

Darling, M., & Steinberg, L. (1993). Parenting style as context: An integrative model.
Psychological Bulletin, l_13, 487-496.

Davies, W. H., Zucker, R. A., Noll, R. N., & Fitzgerald, H. (1989). Early
socialization practices in alcoholic families: The relationship of child-rearing patterns to
demographics and parental psychopathology. Unpublished manuscript, Michigan State University.

DeMarsh, J. , & Kumpfer, K. L. (1986). Family-oriented interventions for the prevention
of chemical dependency in children and adolescents. In S. Griswold-Ezekoye, K. L. Kumpfer,
& W. J. Bukoski (Eds.), Childhood and chemical abuse: Prevention and intervention (pp. 117-
151). New York: Haworth.

Dielman, T. E., Barton, K., & Cattell, R. B. (1977). Relationships among family
attitudes and child-rearing practices. Journal of Gen_etic Psychology, Q9, 105-112.

 

Dishion, T. J., Kavanaugh, K., & Reid, J. B. (1989, November). Child-rearing vs. peer
interventions in the reduction of risk for adolescent substance use and adjustment problems: A
secondary prevention strategy. Paper presented at the Conference for the Advancement of
Applied Behavior Therapy, Washington, DC.

Dishion, T. J ., & Patterson, G. R. (1992). Age effects in parent training outcome.
Behavior Therapy, B3, 719-729.

Duckro, P. N. (1991). Biofeedback in the management of headache: II. Headache
Quarterly, 2, 17-22.

163

Dumas, J. E. (1989). Treating antisocial behavior in children: Child and family
approaches. Clinical Psychology Review, 2, 197-222.

Dumas, J. E. (1986). Parental perception and treatment outcome in families of
aggressive children: A causal model. Behavior Therapy, _11, 420-432.

Dumas, J. E., & Albin, J. (1986). Parent training outcome: Does active parental
involvement matter? Behavior Research and Therapy, a, 227-230.

Dumas, J. E., & Serketich, W. J. (1994). Maternal depressive symptomatology and child
maladjustment: A comparison of three process models. Behavior Therapy, _2_5_, 161-181.

Dumas, J. E., & Wahler, R. G. (1983). Predictors of treatment outcome in parent

training: Mother insularity and socioeconomic disadvantage. Behavioral Assessment, _5_, 301-
313.

Dumas, J. E., & Wahler, R. G. (1985). Indiscriminate mothering as a contextual factor
in aggressive-oppositional child behavior: "Damned if you do and damned if you don’t". Journal

of Abnormal Child Psychology, l_3, 1-17.

 

Dunn, N. J ., Jacob, T., Hummon, N ., and Seilhamer, R. A. (1987). Marital stability
in alcoholic-spouse relationships as a function of drinking pattern and location. Journal of
Abnormal Psychology, ,9_6, 99-107.

Edwards, D. W., Yarvis, R. M., Mueller, D. P., & Langsley, D. G. (1978). Does
patient satisfaction correlate with success? Hospital and Community Psychiatg, Q, 188-190.

Elliott, R., & James, E. (1989). Varieties of client experience in psychotherapy: An
analysis of the literature. Clinical Psvchologv Review. 2, 443-467.

 

Ellwood, L. C. (1980). Effects of alcoholism as a family illness on child behavior and
development. Militag Medicine, 188-192.

Emery, R. E., Hetherington, M., & Dilalla, L. F. (1984). Divorce, children, and social
policy. In H. W. Stevenson & A. E. Siegel (Eds.), Child development research and social
policy. Chicago: University of Chicago Press.

Emery, R. E., & O’Leary, K. D. (1982). Children’s perceptions of marital discord and
behavior problems in boys and girls. Journal of Abnormal Child Psychology, l_O, 11-24.

Eron, L. D., Huesmann, L. R., Dubow, E., Romanoff, R. & Yarrnel, P. W. (1987).
Aggression and its correlates over 22 years. In D. H. Crowell, I. M. Evans & C. R.

O’Donnell (Eds.) Childhood aggression and violence: Sources of inﬂuence, prevention, and
control (pp. 249-262). New York: Plenum Press.

 

Evans, R. L., Smith, Halar, E. M., & Kiolet, CL. (1986). Effect of expectation and
level of adjustment on treatment outcome. Psychological Reports, ﬂ, 936-938.

Evans, R. L., Kiolet, C. L., & Smith (1985). Treatment outcome as a function of

164

expectations and prior adjustment. Journal of Social Behavior and Personalig, _1_, 133-136.

Famularo, R., Stone, K., Barnum, R., & Wharton, R. (1986). Alcoholism and severe
child maltreatment. American Journal of Orthopsychiatry, §_6_, 481-485.

Farrington, D. P. & West, D. J. (1981). The Cambridge study in delinquent
development. In S. A. Mednick & A. E. Baert (Eds.), Prospective longitudinal research: An
empirical basis for the prevention of psychosocial disorders (pp. 296-297). New York: Oxford
University Press.

Fiester, A. R., & Rudestam, K. E. (1975). Multivariate analysis of the early dropout

process. Journal of Consulting and Clinical Psychology, 4_3_, 528-535.

Filstead, W. J ., McElfresh, O., & Anderson, C. (1981). Comparing the family
environments of alcoholic and ’normal’ families. Journal of Alcohol and Drug Education, BB,
24-31.

Fitzgerald, H. E., Sullivan, L. A., Ham, H. P., Zucker, R. A., Bruckel, S., Schneider,
A. M., & Noll, R. B. (1993). Predictors of behavior problems in three-year-old sons of
alcoholics: Early evidence for the onset of risk. Child Development, 6_4, 110-123.

Fitzgerald, H. E., Sullivan, L. A., Bruckel, S., Schneider, A., Zucker, R. A., & Noll,
R. B. (1989). Parental aggression related to behavior problems in three-year-old sons of
alcoholics. Presented at the 4th World Congress, World Association of Infant Psychiatry and
Allied Disciplines, Lugano, Switzerland.

Fitzgerald, H. E., Zucker, R. A., & Yang, H. (1995). Developmental systems theory
and alcoholism: Analyzing patterns of variation in high risk families. Psychology of Addictive
Behaviors, 2, 1-15.

Fleischman, M. J. (1981). A replication of Patterson’s "Intervention for Boys with
Conduct Problems". Journal of Consulting and Clinical Psychology, Q, 342-351.

Fleming, J. P., Kellam, S. G., & Brown, C. H. (1982). Early predictors of age at ﬁrst
use of alcohol, marijuana, and cigarettes. Drug and Alcohol Dapendence, 2, 285-303.

Forehand, R., & Atkeson, B. M. (1977). Generality of treatment effects with parents
as therapists: A review of assessment and implementations procedures. Behavior Therapy, _8_,
575-593.

Forehand, R., Furey, W., & McMahon, R. (1984). The role of maternal distress in a
parent training program to modify child noncompliance. Behavioral Psychotherapy, l_2, 93-108.

Forehand, R., Middlebrook, J., Rogers, T., & Stifﬂe, M. (1983). Dropping out of
parent training. Behavior Research and Therapy, ﬂ. 663-668.

Frank, 8., Hole, C. B., Jacobson, S., Justkowski, R., & Huyck, M. (1986).
Psychological predictors of parents’ sense of competence and control and self- versus child-
focused gratiﬁcations. Developmental Psychology, B2, 348-355.

165

Frieswyk, S. H., Allen, J. G., Colson, D. B., Coyne, L., Gabbard, G. O., Horwitz, L.,
& Newsom, G. (1986). Therapeutic alliance: Its place as a process and outcome variable in
dynamic psychotherapy research. Journal of leting and Clinical Psychology, B4, 32-38.

Frankel, R., & Simmons, J. Q. (1992). Parent behavioral training: Why and when some
parents drop out. Journal of Clinical Child Psychology, 2_1, 322-330.

Frick, P. J., Lahey, B. B., Loeber, R., Stouthamer-Loeber, M., Christ, M. A. G., &
Hanson, K. (1992). Familial risk factors to oppositional deﬁant disorder and conduct disorder:
Parental psychopathology and maternal parenting. Journal of Consulting and Clinical
Psychology, 69, 49-55.

Furey, W. M., & Basili, L. A. (1988). Predicting consumer satisfaction in parent
training for noncompliant children. Bahavior Therapy, 12, 555-564.

Gard, G. C., & Berry, K. K. (1986). Oppositional children: Taming tyrants. Journal
of Clinical Child Psychology, Q, 148-158.

Gardner, F. E. (1987). Positive interaction between mothers and conduct-problem
children: Is there training for harmony as well as ﬁghting? Journal of Abnormal Child
Psychology, Q, 283-292.

Gersten, J. C., Langner, T. S., Eisenberg. J. C., Simcha-Fagan, O. & NcCarthy, E. D.
(1976). Stability and change in types of behavioral disturbance of children and adolescents.
Journal of Abnormal Child Psychology, 4, 111-127.

Ghodsian, M., Fogelman, K., Lambert, L., & Tibbenharn, A. (1980). Changes in
behaviour ratings of a national sample of children. British Journal of Social and Clinigl
Psychology, 12, 247-256.

Gomes-Schwartz, B. (1978). Effective ingredients in psychotherapy: Prediction of
outcome from process variables. Journal of Consulting and Clinical Psychology, 4_6, 1023-1035.

Goodwin, D. W. (1985). Alcoholism and genetics: The sins of the fathers. Archives
of Generaﬂ’svchm, 42, 171-174.

Gottman, J. M., & Katz, L. F. (1989). Effects of marital discord on young children’s
peer interaction and health. Developmental Psychology, 375-381.

Griest, D., & Forehand, R. (1982). How can I get any parent training done with all
these other problems going on? The role of family variables in child behavior therapy. Child

and Family Behavior Therapy, 4, 73-79.

Griest, D., Forehand, R., Rogers, T., Breiner, J., Furey, W., & Williams, C. A. (1982).
Effects of parent enhancement therapy on treatment outcome and generalization of a parent
training program. Behavior Research and Therapy, E, 429-436.

Griest, D. L., Forehand, R., & Wells, K. C. (1981). Follow-up assessment of parent

166

behavioral training: An analysis of who will participate. Child Study Journal, 41, 221-229.

Halikas, J. A., & Rimmer, J. D. (1974). Predictors of multiple drug abuse. Archives
of Gergal Psvchiam, 31, 414-418.

Hamilton, M. (1960). A rating scale for depression. Journal of Neurosurgeg and
Psychiatg, Q, 56—62.

Hansson, L., & Berglund, M. (1987). Individual measures of outcome vs. standardized
rating scales in evaluation of in-hospital psychiatric treatment: A methodological study. Acta
Psychiatrica Scandinavica, 7_5, 275-282.

Harwood, M. & Leonard, K. (1989). Family history of alcoholism, youthful antisocial
behavior and problem drinking among DWI offenders. Journal of Studies on Alcohol, 3, 210-
216.

Hawkins, J. D., Catalano, R. F., & Miller, J. Y. (1992). Risk and protective factors
for alcohol and other drug problems in adolescence and early adulthood: Implications for
substance abuse prevention. Psychological Bulletin, Ll_2_, 64-105.

Heather, N. (1982). Alcohol dependence and problem drinking in Scottish young
offenders. British Journal on Alcohol Ed Alcoholis_r_n, 1_7_, 145-153.

Heineman, S., & Yudin, L. (1974). The consumer and evaluator: Perceptions and
satisfactions level of former clients. Journal of Communig Psychology, _2, 21-23.

Helzer, J. E., & Pryzbeck, T. R. (1988). The co-occurrence of alcoholism and other
psychiatric disorders in the general populations and its impact on treatment. Journal of Studies
on Alcohol, 4_9_, 219-224.

Hesselbrock, V. M., Hesselbrock, M. N ., & Stabenau, J. R. (1985). Alcoholism in men
patients subtyped by family history and antisocial personality. Journal of Studies on Alcohol, 46,
59-64.

Hetherington, E. M., Cox, M., & Cox, R. (1981). Effects of divorce on parents and
children. In M. Lamb (Ed.), Nontraditional families. Hillsdale, NJ: Erlbaum.

Hobbes, S. A., Walle, D. L., & Hammersly, G. A. (1991). The relationship between
child behavior and acceptability of contingency management procedures. Child and Family

Behavior Therapy, 12, 95-102.

Holden, G. W., Lavigne, V. V., & Cameron, A. M. (1990). Probing the continuum of
effectiveness in parent training: Characteristics of parents and preschoolers. Journal of

Consulting and Clinical Psychology, 12, 2-8.

Hops, H., Biglan, A., Sherman, L., Arthur, J., Friedman, L., & Osteen, V. (1987).
Home observations of family interactions of depressed women. Journal of Consulting and
Clinical Psychology, 5_5, 341-346.

167

Horton, L. (1984). The father’s role in behavioral parent training: A review. Journal
of Clinical Child Psychology, Q, 274-279.

 

Home, A. M. (1981). Aggressive behavior in normal and deviant members of intact vs.
mother-only families. Journal of Abnormal Child Psychology, 2, 283-290.

Howard, K., Orlinsky, D., & Perilstein, J. (1976). Contribution of therapists to patients’
experiences in psychotherapy: A components of variance model for analyzing process data.
Journal of Consulting and Clinical Psychology, 44, 520-526.

Jacob, T. (1987). Alcoholism and family interaction: Classiﬁcations resulting from
subgroup analyses and multimethod assessments. In K. Hahlweg & M. J. Goldstein (Eds.),

Understanding major mental disorder: The contribution of family interaction research. New
York: Family Process Press.

Jacob, T., Favorini, A., Meisel, S. S., & Anderson, C. M. (1978). The alcoholic’s
spouse, children and family interactions: Substantive ﬁndings and methodological issues. Journal
of Studies on Alcohol, 3, 1231-1251.

 

Jacob, T., & Leonard, K. (1986). Psychosocial functioning in children of alcoholic
fathers, depressed fathers and control fathers. Journal of Studies on Alcohol, 4_7_, 373-380.

Jacob, T., Ritchey, D., Cvitkovic, J .F., & Blane, H. T. (1981). Communication styles
of alcoholic and nonalcoholic families when drinking and not drinking. Journal of Studies on
Alcohol, 42, 466-482.

 

Jacobson, N. S., & Truax, P. (1991). Clinical signiﬁcance: A statistical approach to
deﬁning meaningful change in psychotherapy research. Journal of Consulting and Clinical

Psychology, 52, 12—19.

Jaffe, P., Wolfe, D., Wilson, 8., & Zak, L. (1986). Similarities in behavioral and
social maladjustment among child victims and witnesses to family violence. American Journal
of Orthopsychiatg, 26, 142-146.

Jansen, R. E., Fitzgerald, H. E., Ham, H. P., & Zucker, R. A. (in press). Pathways
into risk: Temperament and behavior problems in three- to ﬁve-year-old sons of alcoholics.

Alcoholism: Clinical and Experimental Research.

Johnson, P. L. & O’Leary, D. (1987). Parental behavior patterns and conduct disorders
in girls. Journal of Abnormal Child Psychology, _1_§, 573-581.

Johnson, S. M., & Christensen, A. (1975). Multiple criteria follow-up of behavior
modiﬁcation with families. Journal of Abnormal Child Psychology, 2, 135-154.

Johnson, S. M., & Lobitz, G. K. (1974). The personal and marital adjustment of
parents as related to observed child deviance and parenting behaviors. Journal of Abnormal
Child Psychology, 2, 193-207.

168

Johnston, L. D., O’Malley, P. M., & Evelard, L. (1978). Drugs and delinquency: A
search for causal connections. In D. B. Kandel (Ed.), Longitudinal Research on DrugUse.
Washington, DC: Hemisphere Publishing Company.

Jouriles, E. N., Barling, J. & O’Leary, D. K. (1987). Predicting child behavior
problems in maritally violent families. Journal of Abnormal Child Psychology, l_5, 165-173.

Jouriles, E. N., & Farris, A. M. (1992). Effects of marital conﬂict on subsequent
parent-son interactions. Behavior Therapy, 22, 355-374.Kandel, DB. (1982). Epidemiological
and psychosocial perspectives on adolescent drug use. Journal of Child Psvchiatg, 2_1, 328-347.

Jouriles, E. N., Murphy, C. M., & O’Leary, K. D. (1989). Effects of maternal mood
on mother-son interaction patterns. Journal of Abnormal Child Psvcholgy, ﬂ, 513-525.

Jouriles, E. N., Pﬁffner, L. J., & O’Leary, S. G. (1988). Marital conﬂict, parenting,
and toddler conduct problems. Journal of Abnormal Child Psychology, _1_6_, 197-206.

Kandel, D. B., Kessler, R., & Margulies, R. (1978). Antecedents of adolescents’
initiation into stages of drug use: A developmental analysis. In D. B. Kandel (Ed.),

Longitudinal research in drug use: Empirical ﬁndings and methodological issues. Washington,
DC: Hemisphere-Wiley.

Kazdin, A. E. (1981). Acceptability of child treatment techniques: The inﬂuence of
treatment efﬁcacy and adverse side effects. Behavior Therapy, 12, 493-506.

Kazdin, A. E. (1985). Treatment of antisocial behavipr in children. Homewood, IL:
Dorsey Press.

Kazdin, A. E. (1987). Treatment of antisocial behavior in children: Current status and
future directions. Psychological Bulletin, Q2, 187-203.

Kemper, T. D., & Reichler, M. L. (1976). Marital satisfaction and conjugal power as
determinants of intensity and frequency of rewards and punishments administered by parents.
The Journal of Genetic Psychology, m, 221-234.

Kirtner, W. L., & Cartwright, D. S. (1958). Success and failure in client-centered

therapy as a function of initial in-therapy behavior. Journal of Consulting Psychology, 22, 329-
333.

Klein, N. C., Alexander, J. F., & Parsons, B. V. (1977). Impact of family systems
intervention on recidivism and sibling delinquency: A model of primary prevention and program

evaluation. Journal of Consulting and Clinical Psychology, 45_, 469-474.

Klein, F., Adrian, A., & Spevak, M. (1974). Patients evaluate therapists. Archives of
General Psychiatg, 21, 113-116.

Knop, J ., Teasdale, T. W., Schulsinger, P., & Goodwin, D. W. (1985). A prospective
study of young men at high risk for alcoholism: School behavior and achievement. Journal of

169
Studies on Alcohol, 46, 273-278.

Kratcoski, P. (1982). Child abuse and violence against the family. Child Welfare, 61,
435-444.

Krosnick, J. A., & Judd, C. M. (1982). Transitions in social inﬂuence at adolescence:
Who induces cigarette smoking? Developmental Psychology, l_8, 359-368.

Kuczynski, L. (1984). Socialization goals and mother-child interaction: Strategies for
long-term and short-term compliance. Developmental Psychology, 2_(_), 1061-1073.

Lahey, B. B., Piacentini, J. C., McBumett, K., Stone, P., Hartdagen, S. E., & Hynd,
G. (1988). Psychopathology and antisocial behavior in the parents of children with conduct

disorder and hyperactivity. Journal of the American Academy of Child and Adolescent
Psychiatg, 21, 163-170.

Lambert, M. (1989). The individual therapist’s contribution to psychotherapy process
and outcome. Clinical Psychology Review, 2, 469-485.

Lebow, J. (1982). Consumer satisfaction with mental health treatment. Psychological
Bulletin, 21, 244-259.

Leonard, K. E., Bromet, E. J., Parkinson, D. K., Day, N. L., & Ryan, C. M. (1985).
Patterns of alcohol use and physically aggressive behavior in men. Journal of Studies on
Alcohol, ﬂ, 279-282.

 

Lewis, D. O., Moy, E., Jackson, L. D., Aaronson, R., Restifo, N., Serra, S., & Simos,
A. (1985). Biopsychosocial characteristics of children who later murder: A prospective study.

American Journal of Psychiatpy, 142, 1161-1167.

Loeber, R. (1982). The stability of antisocial and delinquent child behavior: A review.
Child Development, 62, 1431-1446.

Loeber, R. (1986). Natural histories of conduct problems, delinquency, and associated
substance use: Evidence for developmental progressions. In B. B. Lahey & A. E. Kazdin
(Eds.), Advances in clinical child psychology (Vol. 10). New York: Plenum Press.

Loeber, R. & Dishion, T. (1983). Early predictors of male delinquency: A review.
Psychological Bulletin, 24, 68-99.

Loeber, R., & Stouthamer-Loeber, M. (1986). Family factors as correlates and
predictors of juvenile conduct problems and delinquency. In M. Tonry & N. Morris, (Eds.),
Crime and justice, vol. 7. Chicago: University of Chicago Press.

Lord, D. B. (1983). Parental alcoholism and the mental health of children: A
bibliography and brief observations. Journal of Alcohol and Drug Education, Q, 1-11.

 

Luborsky, L., McLellan, A. T., Woody, G. E., O’Brien, C. P., & Auerbach, A. (1985).

170

Therapist success and its determinants. Archives of General Psychiatg, 42, 602-611.

MacMurray, J. P., Nessman, D. G., Haviland, M. G., & Anderson, D. L. (1987).
Depressive symptoms and persistence in treatment for alcohol dependence. Journal of Studies
on Alcohol, 42, 277-280.

Maguin, E. T. (1991). Evaluation of a community program for the prevention of conduct
problems among preschool age sons of alcoholic fathers. Doctoral dissertation, Michigan State
University, Department of Psychology, East Lansing, MI.

Maguin, E. T., Zucker, R. A., & Fitzgerald, H. E. (1994). The path to alcohol
problems through conduct problems: A family-based approach to very early intervention with
risk. Journal of Research on Adolescerge, 4, 249-269.

Martin, P. J., & Sterne, A. L. (1975). Prognostic expectations and treatment outcome.
Journal of Consulting and Clinical Psychology, ﬂ, 572-576.

McAuley, R. (1982). Training parents to modify conduct problems in their children.
Journal of Child Psychology and Psychiatry and Allied Disciplines, 2_3_, 335-342.

McCord, J. (1979). Some child-rearing antecedents of criminal behavior in adult men.
Journal of Personalitv and Social Psychology, 3_7, 1477-1486.

McCord, J. (1980). Patterns of deviance. In S. B. Sells, R. Crandall, M. Roff, J.

Strauss & W. Pollin (Eds.), Human functioning in longitudinal persgtive: Studies of normal
and psychopathic mpulations. Baltimore: Williams & Wilkins Company.

McKay, GD. (1991). Parents as role models. In D. Robbins (Ed.), Parent training is

prevention: Preventing alcohol and other drug problems among youth in the family. Rockville,
MD: US. DHHS.

McMahon, R. J ., Forehand, R. L., & Griest, D. L. (1981a). Effects of knowledge of
social learning principles on enhancing treatment outcome and generalization in a parent training

program. Journal of Consulting and Clinical Psychology, 42, 526-532.

McMahon, R. J ., Forehand, R. L., Griest, D. L., & Wells, K. C. (1981b). Who drops
out of treatment during parent behavioral training? Behavioral Counseling Qparterly, 1, 79-85.

McNeill, B. W., May, R. J., & Lee, V. E. (1987). Perceptions of counselor source
characteristics by premature and successful terminators. Journal of Counseling Psychology, 21,
86-89.

Merikangas, K. R., Weissman, M. M., Prusoff, B. A., Pauls, D. L., & Leckman, J. F.
(1985). Depressives with secondary alcoholism: Psychiatric disorders in offspring. Journal of
Studies on Alcohol, 46, 199-204.

Miller, P. M., Smith, G. T., & Goldman, M. S. (1990). Emergence of alcohol
expectancies in childhood: A possible critical period. Journal of Studies on Alcohol, _5_1_, 343-

171
349.

Mills, C. J ., & Noyes, H. L. (1984). Patterns and correlates of initial and subsequent
drug use among adolescents. Journal of Consulting and Clinical Psychology, 22, 231-243.

Moos, R. H., & Billings, A. G. (1982). Children of alcoholics during the recovery
process: Alcoholic and matched control families. Addictive Behaviors, 1, 155-163.

Moos, R. H., Finney, J. W., & Chan, D. A. (1981). The process of recovery from
alcoholism: 1. Comparing alcoholic patients and matched community controls. Journal of Sgdies
on Alcohol, 42, 383-402.

Moos, R.H., Finney, J.W., & Gamble, W. (1982). The process of recovery from
alcoholism: 11. Comparing spouses of alcoholic patients and matched community controls.
Journal of Studies on Alcohol, 43, 888-907.

Moos, R. H., & Moos, B. S. (1981). Family Environment Scale Manual. Palo Alto,
CA: Consulting Psychologists Press.

Moos, R. H. & Moos, B. S. (1984). The process of recovery from alcoholism: III.
Comparing functioning in families of alcoholics and matched control families. Journal of Studies
on Alcohol, 46, 111-118.

Nelson, M. L., & Taylor, C. S. (1992, Aug). Openness to experience and working
alliance: A preliminary report. Paper presented at the lOlst Annual Convention of the American
Psychological Association, Toronto, Canada.

Nelson, R. A., & Borkovec, T. D. (1989). Relationship of client participation and
psychotherapy. Journal of Behavior Therapy and Experimental Psychiatry, E, 155-162.

Noll, R. B., & Zucker, R. A. (1985). The Child Behavior Rating Scale - Preschool
Version. Unpublished instrument, Michigan State University, East Lansing, MI.

Noll, R. B., Zucker, R. A., & Greenberg, G. S. (1990). Identiﬁcation of alcohol by
smell among preschoolers: Evidence for early socialization about drugs occurring in the home.
Child Development, 61, 1520-1527.

Noll, R. B., Zucker, R. A., Weil, C., & Greenberg, G. S. (1984, Aug). Cognitive
factors in substance abuse: The case for early learning. Paper presented at the Annual Meeting
of the American Psychological Association, Toronto, Canada.

Nye, C. L. (1992). Parent, family, and treatment process predictors of child outcome
among alcoholic families in an early intervention program. Masters thesis, Michigan State
University, East Lansing, MI.

Nye, C. L., Zucker, R. A., & Fitzgerald, H. E. (in press). Early intervention in the
path to alcohol problems through conduct problems: Treatment involvement and child behavior

change. Jouma1 of Consulting and Clinical Psychology.

172

Oltmanns, T. F., Broderick, J. E., O’Leary, K. D. (1977). Marital adjustment and the
efﬁciency of behavior therapy with children. Journal of Conaulting and Clinical Psychology, ﬂ,
724-729.

Olweus, D. (1980). Familial and temperamental determinants of aggressive behavior in
adolescent boys: A causal analysis. Developmental Psychology, l_6, 644-660.

Orlinsky, D. E. (1989). Researchers’ images of psychotherapy: Their origins and
inﬂuence on research. Clinical Psychology Review, 2, 413-441.

Orlinsky, D. E., & Howard, K. I. (1980). Gender and psychotherapeutic outcome. In
A. M. Brodsky & R. T. Hare-Mustin (Eds.), Handbook of psychotherapy and behavior change
(Chapter 8). New York: John Wiley & Sons.

Patterson, G. R. (1976). The aggressive child: Victim and architect of a coercive
system. In L. A. Hammerlynck, L. C. Handy, & E. J. Mash (Eds.), Behavior modiﬁcation and
families: Th_eorv and researg (Vol. 1, pp. 267-316). New York: Brunner/Mazel.

Patterson, G. R. (1986). Performance models for antisocial boys. American
Psychologist, 41, 432-444.

Patterson, G. R., & Bank, L. (1987). Some amplifying mechanisms for pathologic
processes in families. In M. Gunnar (Ed.) Minnesota Symmsium on Child Psychology.
Hillsdale, NJ: Lawrence Erlbaum Associates, Inc.

Patterson, G. R., & Chamberlain, P. (1994). A functional analysis of resistance during
parent training therapy. Clinical Psychology: Science and practice, 1, 53-70.

Patterson, G. R., & Chamberlain, P. (1988). Treatment process: A problem at three
levels. In L. C. Wynne (Ed.), The sgte of the art in family therapy research: Conjtroveraies ang
recommendations (pp. 189-223). New York: Family Process Press.

Patterson, G. R., Chamberlain, P., & Reid, J. B. (1982). A comparative evaluation of
a parent-training program. Behavior Therapy, _12, 638-650.

Patterson, G. R., Cobb, J. A., & Ray, R. S. (1973). A social engineering technology
for retraining the families of aggressive boys. In H. B. Adams & I. P. Unikel (Eds.), Issues and
trends in behavior therapy (pp. ). Springﬁeld, IL: Charles C. Thomas.

Patterson, G. R., Dishion, T. J., & Chamberlain, P. (1993). Outcomes and
methodological issues relating to treatment of antisocial children. In T. R. Giles (Ed.),
Handbook of Effective Psychothergy (pp. 43-88). New York: Plenum Press.

Patterson, G. R., & Reid, J. B. (1973). Intervention for families of aggressive boys:
A replication study. Behavior Research and Therapy, 11, 383-394.

Patterson, G. R., & Stoolmiller, M. (in press). Relationship of changes in parental
resistance and discipline practices to future clinical outcomes. Behavior Therapy.

173

Patterson, G. R., & Stouthamer-Loeber, M. (1984). The correlation of family
management practices and delinquency. Child Development, ﬂ, 1299-1307.

Peed, S., Roberts, M ., & Forehand, R. (1977). Evaluation of the effectiveness of a
standardized parent training program in altering the interaction of mothers and their noncompliant
children. Behavior Modiﬁcation, 1, 323-350.

Pekarik, G. (1992). Relationship of clients’ reasons for dropping out of treatment to
outcome and satisfaction. Journal of Clinical Psychology, 48, 91-98.

Popkin, M. H. (1991). Parents as family policymakers and rule setters. In D. Robbins

(Ed.), Parent training is prevention: Preventing alcohol and other drug problems among youth
in the family. Rockville, MD: US DHHS.

Porter, B. & O’Leary, K. D. (1980). Marital discord and childhood behavior problems.
Journal of Abnormal Child Psychology, B, 287-295.

Primakoff, L., Epstein, N ., & Covi, L. (1986). Homework compliance: An
uncontrolled variable in cognitive therapy outcome research. Behavior Therapy, _1_7_, 433-446.

Rachel, J. V., Guess, L. L., Hubbard, R. L., Maisto, S. A., Cavanaugh, E. R.,
Waddell, R., & Benrud, C. H. (1982). Adolescent drinking behavior, Vol. 1: The extent and
nature of adolescent alcohol and drug use: The 1974 and 1978 National Sample Studies.
Research Triangle Park: Research Triangle Institute.

Regier, D. A., Farmer, M. E., Rae, D. S., Locke, B. Z., Keith, S. J., Judd, L. L., &
Goodwin, F. K. (1990). Comorbidity of mental disorders with alcohol and other drug abuse.
Journal of the American Medical Association, 12, 2511-2518.

Reid, J. B., & Patterson, G. R. (1988). The development of antisocial behavior patterns
in childhood and adolescence. Euroman Journal of Personalig,

Reider, E., Zucker, R. A., Maguin, E. T., Noll, R. B., & Fitzgerald, H. E. (1989,
Aug). Alcohol involvement and violence towards children among high risk families. Paper
presented at the 97th Annual Meeting of the American Psychological Association, New Orleans.

Reider, E., Zucker, R. A., Noll, R. B., Maguin, E. T., & Fitzgerald, H. E. (1988,
Aug). Alcohol involvement and family violence in a high risk sample: I. Spousal violence.
Presented at the 96th Annual Meeting of the American Psychological Association, Atlanta, GA.

Rice, L. N., & Wagstaff, A. K. (1967). Client voice quality and expressive style as
indices of productive psychotherapy. Journal of Consulting Psychology, 21, 557-563.

Richman, N ., Stevenson, J. & Graham, P. (1982). Pre-school to school. London:
Academic Press.

Robins, L. N. (1978). Sturdy predictors of adult antisocial behavior: Replications from
longitudinal studies. Psychological Medicine, 8, 611-622.

174

Robins, L. N. (1966). Deviant children grown up. Baltimore: The Williams & Wilkins
Company.

Robins, L. N., West, P. A. & Herjanic, B. L. (1975). Arrests and delinquency in two
generations: A study of black urban families and their children. Journal of Child Psychology
and Psychiatry, 16, 125-140.

Robins, L. N., Helzer, J. H., Croughan, J., & Ratcliff, K. S. (1980). The NIMH
diagnostic interview schedule: Its history, characteristics, and validig. St. Louis, MO:
Washington University School of Medicine.

Roosa, M. W., Sandler, I. N., Beals, J., & Short, J. L. (1988). Risk status of adolescent
children of problem-drinking parents. American Journal of Community Psychology, 16, 225-
239.

Rydelius, P. (1981). Children of alcoholic fathers: A longitudinal prospective study.

In S. S. Mednick & A. E. Baert (Eds.), Prosmtive longitudinal research: An empirical basis
for the prevention of psychosocial disorders. New York: Oxford University Press.

Saltzman, C., Luetgert, M. J ., Roth, C. H., Creaser, J ., & Howard, L. (1976).
Formation of a therapeutic relationship: Experiences during the initial phase of psychotherapy
as predictors of treatment duration and outcome. Journal of Consulting and Clinical Psychology,
ﬂ, 546-555.

Sayger, T. V., Home, A. M., & Glaser, B. A. (1993). Marital satisfaction and social
learning family therapy for child conduct problems: Generalization of treatment effects. Journal

of Marital and Family Therapy, Q, 393-402.

 

Schneider, A. M., Sullivan, L. A., Bruckel, S., Fitzgerald, H. E., Zucker, R. A., &
Noll, R. B. (1989, Sept). The relation of parental depression and self-esteem to behavior
problems in three-year-old sons of alcoholics. Paper presented at the 4th World Congress, World
Association of Infant Psychiatry and Applied Disciplines, Lugano, Switzerland.

Schuckit, M. A., & Sweeney, S. (1987). Substance use and mental health problems
among sons of alcoholics and controls. Journal of Studies on Alcohol, 4_8, 528-534.

Schwartz, D. (1991). Early termination from individual psychotherapy: A follow-up
study of the earliest-leaving patients (and prospective patients). Dissertation Abstracts
International, ﬂ, 2784.

Scovern, A. W., Bukstel, L. H., Kilman, P. R., Lawal, R. A., Busemeyer, J., & Smith,
V. (1980). Effects of parent counseling on the family system. Journal of Counselng
Psychology, 21, 268-275.

Seilhamer, R. A., & Jacob, T. (1990). Family factors and adjustment of children of
alcoholics. In M. Windle & J. Searles (Eds.), Children of alcoholics: A critical review of thg
literature (pp 168-186). New York: Guilford Press.

175

Seixas, J. (1977). Children from alcoholic families. In N. J. Estes & M. E. Heinemann
(Eds.), Alcoholism: Development, consequences and interventions. St. Louis: The CV. Mosby
Company.

Selzer, M. L. (1975). A self-administered Short Michigan Alcoholism Screening Test
(SMAST). Journal of Studies on Alcohol, 26, 117-126.

Shelton, J. L., & Levy, R. L. (1981). Enhancing compliance with behavioral
assignments. In J. L. Shelton & R. L. Levy (Eds.), Behavioral assignments and treatment
compliance. Illinois: Research Press.

Sher, K. J., Walitzer, K. S., Wood, P., & Brent, E. E. (1991). Characteristics of
children of alcoholics: Putative risk factors, substance use and abuse, and psychopathology.

Journal of Abnormal Psychology, 1&1), 427-448.

Smith, G. M., & Fogg, C. P. (1978). Psychological predictors of early use, late use,
and non-use of marijuana among teenage students. In D. B. Kandel (Ed.), Longitudinal research

on drug use: Errmirical ﬁndings and methodological issues (pp 101-113). Washington, DC:
Hemisphere-Wiley.

Speer, D. C. (1992). Clinically signiﬁcant change: Jacobson and Truax (1991) revisited.
Journal of Consulting and Clinical Psychology, @, 402-408. .

Stein, D. M., & Lambert, M. (1984). On the relationship between therapist experience
and psychotherapy outcome. Clinical Ps cholo Review, 4, 1-16.

Steinglass, P. (1980a). A life history model of the alcoholic family. Family Process,
1_9, 211-225.

Steinglass, P. (1980b). Assessing families in their own homes. American Journal of
Psychiatg, ﬂ, 1523-1529.

Steinglass, P. (1981). The alcoholic family at home: Patterns of interaction in dry, wet,
and transitional stages of alcoholism. Archives of General Psychiatry, BB, 578-584.

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

Steinglass, P., Tislenko, L., & Reiss, D. (1985). Stability/instability in the alcoholic
marriage: The interrelationships between course of alcoholism, family process, and marital
outcome. Family Process, 2_4, 365-383.

Stewart, M. A., DeBlois, C. S. & Cummings, C. (1980). Psychiatric disorder in the
parents of hyperactive boys and those with conduct disorder. Journal of Child Psychology and
Psychiatry, 2_1, 283-292.

Stewart, M. A. & Leone, L. (1978). A family study of unsocialized aggressive boys.
Biological Psychiatg, g, 107-117.

176

Stoneman, Z., Brody, G. H., & Burke, M. (1989). Marital quality, depression, and
inconsistent parenting: Relationship with observed mother-child conﬂict. American Journal of
Orthopsychiatg, 62, 105-117.

Stoolmiller, M., Duncan, T., Bank, L., & Patterson, G. R. (1993). Some problems and
solutions in the study of change: Signiﬁcant patterns of client resistance. Journal of Consulting

and Clinical Psychology, 61 , 920-928.

Tolan, P. H., & Lorian, R. P. (1988). Multivariate approaches to the identiﬁcation of
delinquency proneness in adolescent males. American Journal of Comqu Psychology, _16,
547-561.

 

Tracey, T. J. (1989). Client and therapist satisfaction over the course of psychotherapy.
Psychotherapy, &, 177-182.

Tremblay, R. E., McCord, J ., Boileau, H., LeBlanc, M., Gagnon, C., Charlebois, P.,
& Larivee, S. (1990, Nov). The Montreal prevention experiment: School adjustment and self-
reported delinquency after three years of follow-up. Paper presented at the 42nd Annual Meeting
of the American Society of Criminology, Baltimore, MD.

US Department of Health & Human Services (1993). Eighth special report to Copgress
on alcohol and health. Washington, DC.

VanVoorhis, P., Cullen, F. T., Mathers, R. A., & Garner, C. C. (1988). The impact
of family structure and quality of delinquency: A comparative assessment of structural and
functional factors. Criminology, 26, 235-261.

Waas, G. A., & Anderson, G. P. (1991). Outcome expectancy and treatment

acceptability: Perceptions of school-based interventions. Professional Psychology: Research and
Practice, 22, 149-154.

Wadsworth, M. E. J. (1980). Early life events and later behavioral outcomes in a British
longitudinal study. In S. B. Sell, R. Crandall, M. Roff, J. Strauss & W. Pollin (Eds.), Human

functioning in longitudinal perspective: Studies of normal and psychopathic populations.
Baltimore: Williams & Wilkins Company.

Wahler, R. G. & Dumas, J. E. (1987). Family factors in childhood psychology:
Toward a coercion-neglect model. In T. Jacob (Ed.), Family interaction and psychopathology:
Theories, methods, and ﬁndings. New York: Plenum Press.

Walker, E., Downey, G., & Bergman, A. (1989). The effects of parental
psychopathology and maltreatment on child behavior: A test of the diathesis-stress model. Child
Development, 60, 15-24.

Weber, M. D., Graham, J. W., Hansen, W. B., Flay, B. R., & Johnson, C. A. (1989).
Evidence for two paths of alcohol use onset in adolescents. Addictive Behaviors, ﬂ, 399-408.

Webster-Stratton, C. (1985). Predictors of treatment outcome in parent training for

177

conduct disordered children. Behavior Therapy, 16, 223-243.

Wechsler, H., & Thum, D. (1973). Teenage drinking, drug use, and social correlates.
Journal of Studies on Alcohol, 24, 1220-1227.

Weist, M. D., Ollendick, T. H., & Finney, J. W. (1991). Toward the empirical
validation of treatment targets in children. Clinical Psychology Review. 3, 515-538.

Wells, L. E., & Rankin, J. H. (1988). Direct parental controls and delinquency.
Criminology, 26, 263-285.

West, M. O. & Prinz, R. J . (1988). Parental alcoholism and childhood psychopathology.
In S. Chess, A. Thomas, & M. E. Hertzig (Eds.), Annual Progress in Child Psychiatry and
Child Development (pp. 278-314). New York: Brunner/Mazel, Inc.

Whipple, E., Fitzgerald, H. E., & Zucker, R. A. (1995). Parent—child interactions in
alcoholic and non-alcoholic families. American Journal of Orthopsychiatg, 66, 153-159.

Wilkins, W. (1973). Expectancy of therapeutic gain: An empirical and conceptual
critique. Journal of Consulting and Clinical Psychology, 42, 69-77.

Williams, C. N. (1987). Child care practices in alcoholic families. Alcohol Health and
Research World, Summer, 74-77,94.

Winans, T., & Cooker, P. G. (1984). The effects of parent education on the marital
dyad: An examination of systems theory. Journal of Marital and Family Therapy, 10, 423-425.

Windle, M. (1989). Adolescent temperament: Childhood problem precursors and
problem behavior correlates. Paper presented at the Annual Meeting of the American Academy
of Child and Adolescent Psychiatry.

Wolf, M. M. (1978). Social validity: The case for subjective measurement or how
applied behavior analysis is ﬁnding its heart. Journal of A lied Behavior Anal sis, 11, 203-214.

Wolfe, D. A., Jaffe, P., Wilson, S. K. & Zak, L. (1985). Children of battered women:
The relation of child behavior to family violence and maternal stress. Journal of Consulting and
Clinical Psychology, 5_3, 657-665.

Zucker, R. A. (1976). Parental inﬂuences on the drinking patterns of their children. In
M. Greenblatt & M. A. Schuckit (Eds.), Alcoholism problems in women and children. New
York: Grune & Stratton.

Zucker, R. A. (1987). The four alcoholisms: A developmental account of the etiologic
process. In P. C. Rivers (Ed.), Alcohol & addictive behaviors. Lincoln: University of
Nebraska Press.

Zucker, R. A. (1991). Scaling the developmental momentum of alcoholic process via
the Lifetime Alcohol Problems Scores. Alcohol and Alcoholism, 1, 505-510.

178

Zucker, R. A. (in press). Pathways to alcohol problems and alcoholism: A
developmental account of the evidence for multiple alcoholisms and for contextual contributions
to risk. In R. A. Zucker, J. Howard, & G. M. Boyd (Eds.), The development of alcohol
problems: Exploring the biopsychosocial matrix of risk. Rockville, MD: National Institute on
Alcohol Abuse and Alcoholism.

Zucker, R. A., Constance, M. W., Baxter, J. A. & Noll, R. B. (1984, Oct). Differences
in interpersonal and individual psychopathology in young families at high risk for the
development of alcoholism. Paper presented at the Society for Life History Research in
Psychopathology meetings, Johns Hopkins University, Baltimore, MD.

Zucker, R. A., Ellis, D. A., & Fitzgerald, H. E. (1994a). Developmental evidence for
at least two alcoholisms: I. Biopsychosocial variation among pathways into symptomatic
difﬁculty. Annals of th_e New YorkAcademy of Sciences, EB, 134-146.

Zucker, R. A., Ellis, D. A., & Fitzgerald, H. E. (1994b). Other evidence for at least
two alcoholisms: II. The case for lifetime antisociality as a basis of differentiation. Unpublished
manuscript, Michigan State University, East Lansing, MI.

Zucker, R. A., & Fitzgerald, H. E. (1991). Early developmental risk factors and risk
for alcohol problems. Alcohol Health and Research World, 16, 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 & D. J. Cohen (Eds.), Manual of developmental

psychopathology, Volume 2. New York: Wiley.

Zucker, R. A., Maguin, E. T., Noll, R. B., Fitzgerald, H. E., & Klinger, M. T. (1990,
Aug). A prevention program for preschool children of alcoholics: Design and early effects.
Paper presented at the Annual Meeting of the American Psychological Association, Boston, MA.

Zucker, R. A., & Noll, R. B. (1980a). The antisocial behavior checklist. Unpublished
instrument, Michigan State University, Department of Psychology, East Lansing, MI.

Zucker, R. A., & Noll, R.B. (1980b). Drinking and drug history. Unpublished
instrument, Michigan State University, Department of Psychology, East Lansing, MI.

Zucker, R. A., & Noll, R. B. (1987). The interaction of child and environment in the
early development of drug involvement: A far ranging review and a planned very early
intervention. Drugs and Socieg, 1, 1-34.

Zucker, R. A., Noll, R. B., Cruise, K., Kriegler, J., Patricia, W., & Mitchell, C.

(1985). A program for the prevention of conduct disorders: Intervention manual. Unpublished
manuscript, Michigan State University, Department of Psychology. East Lansing, MI.

Zucker, R. A., Noll, R. B., Draznin, T. H., Baxter, J. A., Weil, C. M., Theado, D.
P., Greenberg, G. S., Charlot, C., & Reider, E. (1984a, Apr). The ecology of alcoholic
families: Conceptual framework for the Michigan State University Longitudinal Study.

179

Presented at the National Council on Alcoholism National Alcoholism Forum, Detroit, MI.

Zucker, R. A., Noll, R. B., & Fitzgerald, H. E. (1988). Multiple risk child outreach
program. Unpublished manuscript, Michigan State University, East Lansing, MI.

Zucker, R. A., Noll, R. B., Fitzgerald, H. E., & Maguin, E. T. (1989). The prevention
of conduct problems in young alcoholic families. Unpublished manuscript, Michigan State
University, Departments of Psychology and Pediatrics and Human Development, East Lansing,
MI.

Zucker, R. A., Weil, C. M., Baxter, J. A., & Noll, R. B. (1984b, Oct). Differences
in interpersonal and individual psychopathology in young families at high risk for the
development of alcoholism. Presented at the Society for Life History Research in
Psychopathology meetings: Johns Hopkins University, Baltimore, MD.

“

 

"111111111111