r “W
LIBRARY

Michigan State
University

\M I

 

 

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

ll DATE DUE DATE DUE DATE DUE

. ll asusvgrgflnm
l I ‘

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

MSU Is An Affirmative Action/Equal Opportunity Institution
ammo-9..

 

PARENTAL DEPRESSION AND
SEX DIFFERENCES IN PARENT [SON INTERACTIONS

By

Rosanne du Bois Brouwer

A THESIS

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

MASTER OF ARTS

Department of Psychology

1991

ABSTRACT

PARENTAL DEPRESSION AND
SEX DIFFERENCES IN PARENT/SON INTERACTIONS

By

Rosanne du Bois Brouwcr

Examined the effects of parental depression on parents’ verbalizations to their pre-
school age sons in a structured playroom setting. Subjects were part of a larger
longitudinal study of families with alcoholic fathers. The Dyadic Parent/Child
Interaction Coding System, the Beck Depression Inventory, and the Hamilton
Depression Scale were used to examine the effects of depression on parents’
verbalintions during the play sessions. More depressed mothers were predicted to
display fewer positive verbalizations than less depressed mothers, whereas fathers’
depression was predicted to be unrelated to their play behavior. Contrary to
hypotheses, both mothers pad fathers gave more commands as their depression scores
increased. Higher depression as measured by both the Hamilton and Beck instruments
resulted in fewer questions being asked. Sex of parent affected questions only when
measured by Beck, with more depressed fathers asking significantly fewer questions

than more depressed mothers.

ACKNOWLEDGEMENTS

I am indebted to my advisor, Dr. Hiram E. Fitzgerald, who supported my
efforts during the preparation of this work. I also wish to thank the other members of
my committee, Dr. Robert A. Zucker and Dr. Elaine Donelson for their
encouragement.

I wish to thank Matthew DeYoung, whose cheerful and patient help with the
data analysis was most appreciated.

I also wish to acknowledge the efforts of Dr. Robert B. Noll in incorporating
the Dyadic Parent/Child Interaction Coding System into the Michigan State University
Family Study.

Finally, I wish to acknowledge several individuals who spent many hours
mastering the coding system and coding the family playroom sessions, namely, Hope

Pierson, Cheryl Burda, and Francis Canfield.

This work was supported in part by grants to R.A. Zucker, R.B. N011, and HE.
Fitzgerald from the National Institute on Alcohol Abuse and Alcoholism (AA 07065),

and from the Michigan Department of Mental Health, Prevention Services Unit.

TABLE OF CONTENTS

Page

INTRODUCTION ............................................ 1
Depressed Mothers’ Behavior Towards Their Children ............... 3
Father-Child Interactions and Depression ........................ 7

Sex Differences In Parent Child Relationships ..................... 9
Depression and Alcohol .................................... 11

Other Variables Related to Parenting ........................... 13

The Current Study ........................................ 14
Summary .............................................. 15
HYPOTHESES .............................................. 16
METHOD .................................................. 18
Subjects ............................................... 18

Data Collection Procedures .................................. 19

. Dyadic Parent-Child Interaction Coding System (Eyberg) ........ 19
Depression Measures ...................................... 22
Lifetime Alcohol Problems Scores (LAPS) ....................... 23
RESULTS .................................................. 25
Beck Score Results ....................................... 28
Hamilton Score Results .................................... 29

AN COVA with additional variables ............................ 30
Regression Analyses ...................................... 31
Alcohol Problems Status and Eyberg Categories ................... 32
DISCUSSION ............................................... 34
SUMMARY ................................................. 41

REFERENCES .............................................. 80

iv

LIST OF TABLES

Table 1. Description of behavior categories (Eyberg instrument) .............. 21
Table 2. Sociodemographic information for the sample ................... 27
Table 3. Means and Standard Deviations for the sample ................... 27

AN COVA tables .....

APPENDIXES

 

........................................................................................................... 42
Table 1A. AN COVA with Sex and Hamilton Depression Only ................... 4 2
Table 2A. AN COVA with Sex and Hamilton Depression Only ................... 4 3
Table 3A. AN COVA with Sex and Hamilton Depression Only ................... 44
Table 4A. AN COVA with Sex and Hamilton Depression Only ................... 4 5
Table 5A. AN COVA with Sex and Hamilton Depression Only ................... 46
Table 6A. AN COVA with Sex and Hamilton Depression Only ................... 4 7
Table 7A. AN COVA with Sex and Hamilton Depression Only ................... 4 8
Table 8A. AN COVA with Sex and Hamilton Depression Only ................... 4 9
Table 9A. AN COVA with All Variables Entered. ........................................ 50
Table 10A. AN COVA with All Variables Entered. ........................................ 51
Table 11A. AN COVA with All Variables Entered. ........................................ 52
Table 12A. AN COVA with All Variables Entered. ........................................ 53
Figures ............................................................................................................................... 54
Figure 1A. Relationship Between Beck Depression and Percent Commands

by Sex ............................................................................................ 54
Figure 2A. Relationship Between Beck Depression and Percent Questions by
Sex ................................................................................................. 55

Figure 3A. Relationship Between Hamilton Depression and Percent
Commands by Sex - - - - -- - ----.56

Figure 4A. Relationship Between Hamilton Depression and Percent
Questions by Sex ........................................................................... 57
Stepwise Multiple Regression with All Variables Entered. ............................................. 58
Table 13A. Procedure for Dependent Variable PRAI_PCI‘ ............................ 5 8
Table 14A. Procedure for Dependent Variable COMM_PCI‘ ........................ .59
Table 15A. Procedure for Dependent Variable CRIT_PCI‘ ............................ 61
Table 16A. Procedure for Dependent Variable QUES_PCT ........................... 62

LAPS upper and lower quartile test results ..................................................................... 64

 

 

 

Table 17A. Variable: LAPS .............................................................................. 64
Table 18A. Variable: PRAI_PC1‘ ..................................................................... 64
Table 19A. Variable: COMM_PCT .................................................................. 6 5
Table 20A. Variable: CRIT_PCI‘ ..................................................................... 65
Table 21A. Variable: QUES_PCT .................................................................... 66
Table 22A. Variable: PRAI_PCI‘ ..................................................................... 66
Table 23A. Variable: COMM_PCT .................................................................. 6 7
Table 24A. Variable: CRIT_PCI‘ ..................................................................... 67
Table 25A. Variable: QUES_PCT .................................................................... 68
Table 26A. Variable: LAPS .............................................................................. 69
Table 27A. Variable: PRAI_PCT ..................................................................... 69
Table 28A. Variable: COMM_PCT .................................................................. 7 0
Table 29A. Variable: CRIT_PCI‘ ..................................................................... 70
Table 30A. Variable: QUES_PCI‘ .................................................................... 71
Table 31A. Variable: PRAI_PCI‘ - -- -- ........ - - ....;71
Table 32A. Variable: COMM_PCT .................................................................. 7 2
Table 33A. Variable: CRIT_PCI‘-- - -- -- - .......... -72
Table 34A. Variable: QUES_PCT .................................................................... 73
Depression score distribution plots ................................................................................... 74
Table 35A. Hamilton Depression Score Distribution ...................................... 274
Table 36A. Normal Probability Plot for Hamilton Depression Scores ........... 74
Table 37A. Beck Depression Scores Distribution ............................................ 7 5
Table 38A. Normal Probability Plot Beck Depression Inventory Score ......... 7 6
Informed consent form ...................................................................................................... 77
Data recording sheet.-- "3.....79

vii

INTRODUCTION

The purpose of this study was to examine the effects of parental depression on
the interactions between parents and their pre-school age sons in a structured playroom
setting. Of particular interest were the possible differences between depressed mothers
and depressed fathers in this setting. Since the sample consisted of families with
alcoholic fathers, the relationship between alcohol use and depression was also of
interest.

Depression is a relatively common psychiatric disorder, with women two to
three-times more likely to experience depression than men (Weissman & Klerman,
1977; Hopkins, Marcus, & Campbell, 1984). Women who live in financially strained
circumstances and who have responsibility for young children are more likely than
Other women to become depressed (Brown, Bhrolchain, & Harris, 1975; Pearlin &
Johnson, 1977; Radloff, 1975; Goldman & Ravid, 1980).

Patterson (1981) cites evidence from a survey in My (April 14,
1976) in which 23% of housewives interviewed reported tension, headaches, and
depression. Patterson believes this is often attributable to the high rates of aversive
events that mothers are exposed to, and that these events are most often associated

with children. Fawl (1963) reported data showing that in non-problem, middle-class

2

homes, an average of 3.4 disturbances per hour occurred between mothers and
preschool children. Furthermore, the median number of these disturbances from two-
and three-year-olds was more than twice that from four- and five-year-olds. Patterson
(1981) points out that during this time of frequent stress, mothering of young children
requires a high response rate, and mothers rarely receive a comparable rate of positive
reinforcement for their responsivity. Since it is reasonable to assume that mothers in
non-distressed families have better support systems for the rearing of young children
than do mothers in distressed families, the symptoms of maternal stress and depression
should be higher in distressed families than in non-distressed families. Collette (1983)
reported that, in a sample of young mothers, levels of depression increased in relation
to the amount of stress, and decreased when the mother was involved in a supportive
social network.

There is evidence that children of depressed mothers are at risk for emotional
illness (Beardslee, Bemporad, Keller, & Klerman, 1983; Weissman, Paykel, &
Klerman, 197 2; Rutter, 1966; see also review by Downey & Coyne, 1990), behavior
problems (Dumas, Gibson, & Albin, 1989; Cox, Puckering, Pound, & Mills, 1987;
Forehand, Lautenschlager, Faust, & Graziano, 1986; Williams & Carmichael, 1985),
and cognitive deficits (Gruenebaum, Cohler, Kaufi‘man, & Gallant, 1978). In a sample
of urban working class mothers, children of depressed mothers were found to have
more emotional and behavioral problems than children of non-depressed control
mothers. Children of depressed mothers had eating difficulties, problems in

relationships with peers or parents, and poor attention combined with overactivity

3
when compared with children of non-depressed mothers (Cox et al., 1987). Maternal

depression has also been linked to the development of childhood depression (McKnew
& Cytryn, 1973; Orvaschel, Weissman, &,Kidd, 1980). Dumas, Gibson, and Albin
(1989) suggest that the relationship between maternal depression and childhood
depression is robust because it has been demonstrated in studies where maternal
depression levels were mild to moderate (e.g., Forehand & Brody, 1985) as well as in

studies where mothers had major affective disorders (e.g., Weissman & Paykel, 1974).

De 8 M ’BhviorTowardsTheirChil n

It is important to clarify at the outset that definitions of depression differ in
studies related to maternal depression and child behavior. Most studies focus on clinic
referred children with behavior problems and discuss maternal self-report depression
symptoms as they relate to the mothers’ descriptions of their child’s behavior. Many
of these studies do not separate clinically diagnosed depression from depressive
symptoms or depressed mood. In fact, some studies seem to use the terms
interchangeably (Fergusson, Horwood, Gretton, & Shannon, 1985). Other work uses
only clinically diagnosed mothers, notably Weisman and Paykel (1974) and
Longfellow, et. al (1982).

Very few of the subjects in the current study reach levels of clinical depression,
either with Beck self-report data or Hamilton clinician rated depression. It may be
important to distinguish between clinically diagnosed depression, depression

symptoms, depressive mood, and transient stress-related depression or "bad mood"

4

Because subjects in the current study are not at clinical levels of depression, results in
this study should be viewed as reflecting the behavior of parents as their depression
level increases, and not as reflective of the behavior of a clinically depressed
population.

Some studies show depressed mothers to be less effective parents. Weissman
et al. (1972) found that depressed mothers were unable to be normally involved in
their children’s lives, unable to be positive models for their children, withdrew
physically and emotionally from their children, and complained about being intolerant
of their children’s noise. Others have found depressed mothers to be inconsistent in
their use of discipline (Susman, Trickett, Ianotti, Hollenbeck, & Zahn-Waxler, 1985;
Zelkowitz, 1982). Maternal depression has been associated with verbal and physical
punitiveness and less nurturance. Depressed mothers were much more likely to use
hostile and dominating styles, both in initiating behavior to their children, and when
responding to children’s requests (Longfellow, Zelkowitz, & Saunders, 1982) . They
typically yelled and threatened. Depressed mothers were much less likely to use styles
that conveyed warmth, affection, or positive affect. Attentiveness, responsiveness, and
warmth were more likely in families with non-depressed mothers (Longfellow, et al.,
1982; Weissman & Paykel, 1974; Rutter, 1966). Longfellow et al. (1982), whose
observational study used a sample of low-income mothers, found that the severity of
mothers’ depression was related to a style of parenting that consisted of more
warnings and prohibitions to children, greater use of hostile and dominating control

techniques, and less compliance with their children’s requests. Effects of depression

5
on maternal behavior were present regardless of other important background

characteristics of the family (Longfellow et al., 1982).

Depressed mothers are less likely than normal mothers to meet their pro-school
age child’s demands for attention and communication. Longfellow et al. (1982) found
that depressed mothers were unresponsive to their child’s dependency needs; and their
child’s bids for attention, help, and food tended to be ignored or unfilled. Depressed
mothers were also less likely to initiate nurturant interactions, and more likely to
initiate interactions that encouraged prosocial behavior, i. e., by issuing warnings,
reminders, and commands (Longfellow, et al., 1982).

Self report measures corroborate observational data regarding the behavior of
depressed women and their children. Depressed women report less tolerance for
dependency seeking behaviors in their children (Longfellow, et al., 1982). Zelkowitz
(1982) found that mothers with high depression levels tended to demand immediate
compliance from their children. Depressed women also reported "considerable"
resentment towards their children, whereas non-depressed women did not (Weissman,
et al., 1972). Weissman & Paykel (1974) also found that depressed mothers report
that they experience friction, impaired communication, lack of affection, and
resentment toward their children.

Some depressed mothers tend to get overinvolved in their children’s lives.
Weissman (1974) found that depressed mothers were overinvolved in their children’s
lives and acted in overprotective ways. According to Weissman (1974), "some

depressed women engaged in a frenzied pursuit of activities with their children.

6

However, though this involvement was interpreted by less intimate associates as
‘togetherness,’ closer observation showed that such over involvement actually diluted
emotional closeness and appeared to compensate for a sense of guilt and deeper
disinterest" (p. 113). For example, depressed mothers may be involved with PTA, car
pools, field trips, act as lunch room mothers, etc., but individual contacts with their
children are cold and distant. Moreover, excessive involvement with children’s
activities was sometimes associated with verbal and physical abuse of the children
(Weissman, 1974).

When mothers were asked about changes in their child’s behavior when the
mothers themselves were depressed, only 12% reported there was no change. The
most common pattern observed by Cox et al. (1987) and shown by more than half the
children, was one in which the child approached the mother, clung to her, and
comforted her. Cox reports it was a ”very moving sight" to watch two-year-olds
comforting their distressed mothers (p.921).

There is some conflicting evidence in the literatme regarding the effects of
maternal depression behavior on children. Rogers and Forehand (1983) found no
relationship between maternal depression and maternal behavior towards children.
Forehand and Brody (1985) found maternal depression to be related to maternal
perceptions of child maladjustment but not to maternal behaviors. Forehand,
Lautenschlager, Faust, & Graziano (1986), indicated that there were significant paths
from maternal depression to maternal perceptions of child maladjustment, to increased

commands, and to increased child non-compliance. But when Dumas et al., (1989)

7

controlled for maternal perception, the opposite was found. High levels of maternal
depression were associated with high levels of compliance.

These confusing and conflicting results reflect possible methodological
problems. Webster-Stratton & Hammond (1988) report that studies of depressed
mothers have been hampered by such problems as small sample sizes, lack of
appropriate control groups, and reliance on data from lower-class, clinic referred

mothers.

F r- il In tions d ssion

Much less has been written about father-child interactions. Until quite recently,
the study of infant and preschool development has focused almost entirely on the
mother, whether in psychoanalytic theory, attachment theory, or social learning theory.
Only in the last 20 years have psychologists begun to study father-child interactions
seriously and systematically. The fact that many families now have two working
parents has not increased the actual amount of time fathers spend in caregiving
(Nyquist, Slivken, Spence, & Helrnreich, 1985). Indeed, mothers spend more time in
caregiving functions compared to fathers, even when the mothers also hold full time
jobs outside the home (Kotelchuck, 1976; Nyquist, et al., 1985). However, the issue
of fathers as nurturers has become the focus for some researchers.

Regardless of amount of time spent, there is evidence suggesting that fathers
have a different way of interacting with their children than mothers. At as early as 4

months of age, fathers have been observed to play more games than mothers during

8

interactions with infants, as well as show higher levels of affect (Y ogman, 1982).
Yogman describes the difference this way: "...interaction with fathers can be
characterized as heightened and playful whereas the interactions with mothers
appeared more smoothly modulated and contained" (p.250).

Lamb (1981) suggests that the mother’s biological connection to the child leads
to her role as primary caregiver, thereby freeing fathers to focus on play and
connection to the outside world. Yogman (1982) writes that historically the general
psychological function of fathers has been "the fostering of autonomy and the
enhancement of individuation" (p. 223). Whether this notion is based in biology or
not is an open question (Clarke—Stewart, 1980), but it seems apparent that role
expectations play a large part in the different quality of parenting by mothers and
fathers (Clarke-Stewart, 1980; Yogman, 1982).

Bright and Stockdale (1984) found that fathers controlled and directed their
children during a play session more than mothers, and mothers were more quiet than
fathers during play with their children. Bright and Stockdale suggest this may be
because fathers, and indeed society in general, have the expectation that it is the
father’s role in child-rearing to teach instrumental skills. In their study, children also
controlled and directed their fathers more than they did their mothers, supporting the
idea that father-child interactions are more instrumental in nanne. However, the
research regarding father-child relationships and involvement is limited because most
of the work has been conducted in "traditional" nonemployed-mother families. With

the current figures in the United States at more than 50% of mothers of preschoolers

9
working outside the home (Hayghe, 1990), this pattern is not representative of a

significant number of contemporary American families. More research is needed on

the changes that may or may not be taking place as roles change.

Sex Diffgenggs In Pgent Qild Relag'gnships

There is some support in the literature for a sex difference in roles for mothers
and fathers regarding the nature of the relationship with their children. Clarke-
Stewart (1978, 1980) reported that in interactions with 30-month-old children, fathers
were more likely to reward children drning play and were rated higher than mothers in
their ability to engage the child in play and in the fathers’ own enjoyment and
involvement in the activity. Easterbrooks and Goldberg (1984) obtained similar
results. They found that whether fathers take responsibility for caregiving is less
strongly related to variations in child development than is the amount of time they
spend with the child, supporting the idea that the central role of the father may be
playmate rather than caregiver. Kotelchuck (197 6) has presented evidence from
parental interviews that father-child interaction is more likely to involve play than is
mother-child interaction, whereas mother-child interaction is more likely to involve
caregiving. Lamb (1976) has presented observational data showing fathers were more
likely to hold infants during play, while mothers held them for caregiving.

Patterson (1981) reported that fathers have a high rate of play in both
distressed and non-distressed families, suggesting that fathers are less likely to change

their behavior with their children based on personal stresses. Rather, Patterson argues

10

that mothers are culturally assigned the role of crisis manager in distressed families,
and fathers, who engaged in considerably more normative behavior (such as reading
the newspaper), might be more appropriately labeled as a "guest". Patterson (197 6)
has reported elsewhere that fathers of clinically referred problem children had MMPI
scores indistinguishable from fathers in non-distressed families. Fathers seem to be
able to escape more effectively from the stresses of family life through work and other
activities that allow them to remove themselves, either physically or psychologically
and emotionally. Since the sample in the current study consists of all alcoholic
fathers, it is important to note that this role difference exists in distressed families as
well as non-distressed families. -

There has been virtually no research in which depressed fathers were actually
observed interacting with their children. One study by Webster-Stratton (1988)
examined mothers’ and fathers’ perceptions of child deviance using observational data.
She also used a measure of parenting stress (Abidin, 1983) which includes depression
as one of the eight subscales. Mothers and fathers in her study also completed the
Beck Depression Inventory. For mothers, both Beck depression score and level of
parenting stress were correlated with negative behaviors toward the child. For fathers,
these measures were not correlated with behavior towards the child. These data
support the hypotheses of the current study. They suggest that mothers’ distress level
and depression are related to the way they interact with their children but for fathers,

these personal adjustment measures do not affect parenting behavior.

l 1
It may be that depressed and distressed mothers have lower tolerance for child

misbehavior. It may also be true that fathers use other avenues to cope with
depression and stress.

Downey & Coyne (1990), in their review of the research on children of
depressed parents, did not find any study in which depressed fathers were observed
with their children. The bias present in parent-child research toward the mother as
much more important in child development than the father has left a large gap in the
literature. One advantage of the current study is the use of direct observation to

measure father and mother behavior separately in play interactions with their child.

WM

Depression is a frequent correlate of alcoholism (West & Prinz, 1987). In an
investigation of comorbidity of mental disorders with alcohol abuse, (Regier, Farmer,
Rae, Locke, Keith, Judd and Goodwin; 1990) affective disorders were found in 13.4%
of the subjects with an alcohol disorder, compared with 7.5% among subjects without
an alcohol disorder. Research concerning the relationship of alcohol and depression
that has addressed its effects has mostly focused on child outcomes (see review by
West & Prinz, 1987) rather than actual specific behaviors of mothers and fathers who
are alcoholic or depressed or both. The current study sought to identify whether or
not depression and alcohol problems affect the way mothers and fathers play with their

children.

12

Recent discussion in the literature has focused on whether fathers’ alcoholism .
directly affects child outcomes, or whether the effect on children is mediated by the
wives of the alcoholics (Fitzgerald, Sullivan, Ham, Zucker, Bruckel, Schneider, &
Noll, under review). Fitzgerald et al., using data from the same data set as the current
study, found that maternal variables, including both depression level and mothers’
alcohol problems as measured by the Lifetime Alcohol Problems Scale (LAPS)
(Zucker, 1991), predicted the fathers’ perceptions of their childrens’ problems.

In the current study, the subjects are all families with an alcoholic father.
However, because alcohol problem levels are elevated for both mothers and fathers
when compared to control subjects in this sample (Fitzgerald, et. al, under review),
and because depression reports are similar between mothers and fathers in the subjects
of this study, one possible reason for differences in behavior in the playroom is
gender. It is of course possible that alcoholism and depression affect fathers and
mothers differently as regards their parenting behavior. However, since fathers’
alcoholism may affect children indirectly through the mothers, and since the fathers’
role is that of playmate, it is the prediction of the current study that alcohol problems
status (LAPS) in fathers will not affect their verbalizations to their children. It is
further predicted that depression in mothers will affect their interactions with their
children, while fathers’ behavior in a play situation will not A be affected by their
depression level. Support for this notion also comes from Jacob and Leonard (1986)
who found that children of both alcoholic and depressed fathers had more behavior

problems than the children of normal fathers, but they found no differences between

13
the alcoholic and depressed father groups. Although Jacob and Leonard’s study

reported on child outcomes in alcoholic and depressed father groups and the current
study reflects only on the parents’ verbalizations toward the child, it is reasonable to
suggest that an inability to distinguish between children of alcoholic and depressed
fathers may reflect similar parenting in both alcoholic and depressed fathers. Once
again it should be noted, however, that the subjects in the Jacob and Leonard study
(1986) reached diagnostic criteria for depression. Nevertheless, it seems reasonable
that these two variables are less likely to influence a father’s behavior in a play

situation than a mother’s behavior.

r ' 1 Rela P n '

It is possible that other variables besides depression and alcohol problems may
be contributing to the way parents behave with their children in the playroom. These
include education level of the parents, and following from that, the occupation and
income level of the parents. The families in this study tend to be lower income
families and most are high school graduates only. It is possible that this socio-
economic status may be reflected in the playroom behaviors of the parents.

The number of children in a given family and its differential demands on
parents may also influence the way parents behave when they are given the
opportunity to play undistracted with one child.

Because the MSU Family Study has this information available, these variables

were included in the analyses on an exploratory basis.

14
Th nt S

The current study focuses on the question of sex differences in parents’
behaviors toward their sons in a play situation, particularly as such behaviors relate to
levels of parental depression. To measure parents’ verbalizations in the playroom, the
MSU Family Study uses the Dyadic Parent/Child Interaction Coding System (DPICS)
(Eyberg & Robinson, 1983) to code different types of speech. The DPICS was
deveIOped to allow the clinician to observe parent-child interactions in a playroom
setting. The instrument was designed to identify problem families and to understand
their dynamics. Robinson & Eyberg (1981), found few and relatively small
differences between mothers and fathers in their interactions with young children. The
current study hypothesizes that these differences may be larger for depressed vs. non-
depressed parents.

Use of DPICS has revealed some sex difl'erences in family interaction patterns.
For example, in one study mothers consistently gave more total praise than fathers,
whereas fathers tended to be more directive, and to give more indirect commands
(Robinson & Eyberg, 1981). Another study has demonstrated differences in maternal
verbal behavior between neglect and non-neglect families (Aragona & Eyberg, 1981).
In a group of 27 mother-child dyads, significant findings were reported for direct
commands, verbal praise, verbal acknowledgement, and critical statements, with

negligent mothers consistently lower than controls (Aragona & Eyberg, 1981).

' 15
Summag

In summary, the literature suggests the following conclusions with respect to
the effects of parental depression on how parents behave with their children during
play.

1. Mothers with higher levels of depression are likely to display fewer positive
behaviors toward young children than less depressed mothers.

2. Fathers, depressed or not, tend to view time with young children as
playtime, and do not change their behavior with children in relation to their own
distress level.

3. The effects of an alcoholic father on a child may be mediated through the

child’s mother in her role as caregiver, particularly in younger children.

HYPOTHESES

Since depressed mothers have been reported to resent their children, and to
express anger and resentment over their isolation in the child rearing role, it is
hypothesized that: 1) mothers with higher levels of depression will give less praise to
their children than non-depressed mothers, 2) mothers with higher levels of depression
will give less praise than will depressed or non-depressed fathers, 3) depressed
mothers will use more commands and 4) more critical statements than will fathers, 5)
depressed mothers will use fewer questions on the Eyberg index than will non-
depressed mothers.

Fathers, on the other band, who have not been reporwd to resent their children
or view them as restrictive of their lifestyle, will display more playful and positive
behaviors with their children in the playroom setting. If research suggesting that
fathers view time with their children as pleasurable, and the father’s role as that of
playmate is correct, then fathers will show more positive behavior toward their
children in a play situation than will depressed mothers. 5) Fathers will give more
praise, 6) use fewer commands, 7) make fewer critical statements, and 8) ask more

questions than will depressed mothers.

16

17
Finally, fathers’ LAPS scores will not predict their behavior in the playroom

because fathers tend not to change their behavior with their children based on their
own personal stresses. Mothers’ LAPS scores, similar to their depression scores, will
predict behavior towards their children, with higher LAPS scores predicting less

praise, fewer questions, and more critical statements and commands.

METHOD

Mutts

Subjects are 81 families participating in the Michigan State University Family
Study (Zucker, Noll, & Fitzgerald, 1986), a longitudinal study examining the factors
that may influence child development and health in the offspring of alcoholic and non-
alcoholic parents. The current study uses data from the alcoholic sample only.

Subjects are recruited from local district courts using criteria of 1) men who
are arrested for drunk driving with a blood alcohol concentration (BAC) of at least .15
(or .12 if it is a second offense), 2) at the time of contact are in intact families, and 3)
have biological sons between 3.0 and 6.0 years of age. All candidates are asked by
their probation officers to give their permission to be contacted by project personnel.
Families who give such permission are visited by the project coordinator and are
recruited into the project. The project coordinator visits each family, explains the
study, and secures informed consent for participation (see pages '77 and 78). All

families in the study are paid for their participation.

18

l9
Dag lelectign Procedures
'P nt-ChilIn actinCodin sm ber

Video taped interactions of each parent separately with the child are coded
using the Dyadic Parent Child Interaction Coding System (Eyberg & Robinson, 1983).
Parents and the target child come to the video laboratory at Michigan State University.
Based on a coin flip, one parent is escorted to the playroom with the child while the
other parent is in another room completing questionnaires as part of the larger data
collection protocol. The playroom contains several toys, such as crayons and paper,
Play Doh, trucks, and Lincoln Logs. The playroom has a one-way viewing window
through which video tapes are made of the playroom interactions. After the parent
and child are seated in the playroom, a graduate assistant gives the following
instructions:

There are three 10-minute time periods. We will knock on the window to alert

you to the time changes. The first ten minutes it is your child’s turn to choose

what to play with, and you (parent) should play along with him. The second

ten minutes it is your (parent’s) turn to choose the activity, and your child

should play along with you. The third ten minutes is clean-up time, and I’d

like you to ask your child to clean up. If you finish before the third ten

minutes is up, just say, "We’re done," and I will come and get you.
When this thirty minute protocol is finished, the same procedure is repeated with the

other parent. Both parent and child know they are being video taped.

20
The Eyberg coding system (DPICS) was designed to: a) provide an

observational measure of parent and child behavior as an adjunct to full psychological
evaluation of childhood disorders or parenting skills; b) serve as a pre-treatrnent
assessment of behavior occurring in dyadic family interaction; c) provide a measure of
ongoing progress during therapy that focuses on changing general parent-child
interaction patterns; and d) serve as a behavioral observation measure of treatment
outcome (Robinson & Eyberg, 1984, p. 122).

Observational coding is done live and continuously by graduate students behind
the one-way window, and results in a record of the total frequency of each behavior
during each ten-minute period. The Data Recording Sheet (page 79) is used by
observers to record all data. Verbalizations are coded by making a tally mark in the
appropriate space on the recording sheet each time the particular verbal behavior
occurs (Robinson & Eyberg 1984).

Each verbal behavior is operationally defined in a detailed coding manual
(Eyberg & Robinson, 1983). For each category, the manual provides a general
definition and a series of examples. It also lists specific guidelines to aid in
discrimination between categories. Finally, the manual lists decision rules to use when
there is uncertainty between categories. A brief description of the parental behaviors
is presented in Table 1. Direct and Indirect commands were combined for purposes of
this study because both reflect an instrumental approach by the parent and there were

insufficient numbers to analyze them separately.

21

A number of studies have documented the reliability of the coding categories

(Aragona & Eyberg, 1981; Robinson & Eyberg, 1981). Robinson & Eyberg (1981)

reported interrater reliability based on Pearson correlations of frequency of each

behavior recorded by two observers during 244 five-minute observations. The mean

reliability coefficient for parent behaviors was .91 (range = .76-1.0).

Table l. mscfiptjon Qf Behavig Categgg'gs

Parental Behavior

Description

 

Direct Command

Indirect Command

Descriptive

Reflective
Dex/Reflective
Question
Acknowledgement

Praise

Critical statement

A direct, clearly stated order or direction in declarative form
(e.g., Put the block on the tower.)

An indirect suggestion or question-command which attempts to
lead the child’s verbal or physical activity (e.g., Will you put the
block on the tower?)

Statements, both declarative and question form, which describe
and follow the ongoing activity but contain no evaluative
comment (e.g. You put the block on the tower.)

Statements, both declarative and question form, which exactly
reflect a .verbalization by the child.

Any question asked of the child, may be made a question by an
inflection or a tag (e.g., huh?).

Verbalizations which contain no manifest content (e.g., Okay or
oh.)

A statement indicating approval (e.g., Good, that’s a good place
to put the block.

A statement indicating disapproval (e.g., That’s not a good place
to put the block.)

Adapted from Aragona, J. & Eyberg, S. (1981)

 

22

Subjects consisted of 42 families: a conduct problem group of 20, and a normative
control group of 22, with correlations for each group in this range. In the Aragona &
Eyberg study (1981), reliability coefficients, obtained for eight available families, were
computed using Pearson product-moment correlations between observers for each
behavioral category. For all categories, these coefficients ranged from .65 to 1.00,
with only Parent-Directed Interaction--Acknowledgement falling below .85 (Aragona
& Eyberg, p. 598).

For the current study, inter-rater reliabilities were calculated for each category
individually, based on 42 separate coding periods, using data from three coders.
Pearson correlations for these individual categories ranged fiom .78 to .97. The
correlations for the specific categories used in this study were: Commands: Direct=.80,
Indirect=.78; Praise: labeled Praise (n = only 6) .48, Unlabeled Praise: .80; Critical

statements: .68; and Questions: .97.

W

Self reported depression was measured by the Short Form of the Beck
Depression Inventory (Beck & Beck, 197 2). This inventory requires respondents to
complete 13 statements by picking the answer that best reflects how they feel on that
day. Undergraduates administer the Beck self report measure to each parent at an
earlier data collection session held in the family’s home.

A review of the psychometric properties of the Beck (Beck, Steer, & Garbin,

1988) scanned internal consistency of the long form of the Beck in at least 25 studies.

23
In psychiatric populations, the mean coefficient alpha was .86, with a range of .76 to

.95. In nonpsychiatric samples, the mean coefficient alpha was .81, with a range of
.73 to .92. Beck et al. (1988) claim comparable internal consistency for the short
form. Validity examinations of the Beck Depression Inventory have shown impressive
results in content, concurrent, discriminant, and construct validity (see review by Beck,
et. a1, 1988).

The Hamilton Rating Scale for Depression (HRSD) (Hamilton, 1960), was used
for rating levels of clinical depression. A clinical graduate student rates the level of
clinical depression based on information from an extensive diagnostic interview
conducted in the subject’s home. Subjects are given a current Hamilton depression
score and a "worst ever" score. The I-IRSD covers behavioral, affective, somatic, and
psychological elements of depression. The analyses in this study used current levels
of depression because they were more closely related in time to the status of the
subject at the time of the playroom session. Interrater reliabilities for the HRSD range
from .80 to .90 (Hamilton, 1969).

Lif 'm Al h l lem

As part of the larger longitudinal study at Michigan State University, subjects
are given a measure of lifetime problems related to alcohol. The Lifetime Alcohol
Problems Score (LAPS) (Zucker, 1991) was designed to provide a way of "scaling
differences in extent of alcohol related difficulty over a lifetime." The measure is

based on three aspects of functioning. First is the mac! of alcohol problems, based

24
on how early in the life course alcohol problems begin. Second is the m of

different alcohol related problems. This factor is based on a count of a large number
of different types of alcohol related problems an individual may have experienced at
any time during his or her life. Third is a component which measures the invgiveness
of alcohol problems over the lifetime. This aspect is based on both the rate of
problem display per unit of time, and the duration of the problem (Zucker, 1991).
Each subscale is standardized. The LAPS score is calculated by adding the three
subscores together. Scores for this measure are standardized separately for males and

females within the MSU Family Study sample.

W

The Michigan State University Family Study uses the Revised Duncan (1981)
based on the work of Stevens & Featherrnan (1981), to assess socieconomic status of
the families. This occupation based measure of SES was selected after an extensive
review of the strengths and weaknesses of several possible SES indicators (Mueller &

Parcel, 1981).

RESULTS

The major issue in this study was whether or not current depressive experience
differentially affects the way mothers and fathers speak to their young sons in a
structured playroom setting. Mothers spend more time with their children overall than
fathers, and are more often depressed than fathers. Mothers also tend to relate their
depression to the demands of children, something fathers are less likely to do.
Therefore, it seemed likely that mothers who were more depressed would be likely to
give less praise to their children while playing with them and to ask fewer questions, a
measure of engaging the child in the play. Fathers, on the other hand, less frequently
have the opportunity to play with their children and feel less burdened by the demands
of children. Fathers are likely to view the playroom session as a "luxury" while
mothers are likely to view it as ordinary and even demanding.

While commands may be seen as an instrumental interaction style usually
thought of as the domain of fathers, commands can also be viewed as behavior that
views the child negatively. Commands tend to lessen engagement in the interaction,
with the child merely carrying out the wishes of the parent and not being a participant
in the play. Based on this view, depressed mothers would be more likely to issue

commands to their children than fathers, depressed or not. Since fathers experience

25

26
play time with children as unusual and perhaps even indulgent, fathers would be more

likely to use more engaging styles of interaction than commands. Following this same
logic, depressed mothers would be more likely to be critical of their children than non-
depressed mothers, or fathers, depressed or not.

Analysis of Covariance (AN COVA) was used to analyze the effects of sex and
depression on each of four dependent variables from the Eyberg codes: Praise,
Commands, Critical Statements, and Questions. The four dependent variables were
calculated as a percent of the total number of verbalizations spoken by the parent. A
separate AN COVA was conducted for each of the dependent variables. Because
depression scores did not have a distribution that allowed a split into high and low
depression groups (see plots page 74), depression was treated as a control variable and
allowed to covary with the dependent variable. Separate ANCOVA analyses were
conducted for each of the depression measures.

Table 2 summarizes the demographic characteristics of the sample for the
current study. Table 3 shows the means and standard deviations for mothers and

fathers separately for each of the variables.

27

 

 

 

 

T l 2. ' m hi ' ' f Sam le
Family Variable Mean Range
Family Income $18,000 $4,000-S30,000
Child Variable Mean Range
Child Age
4.43 3.0 - 6.85

Parent Variables Mothers Fathers

Mean Range Mean Range
Age (years) 28.96 21-40 30.57 22-48
Education (years) 12.54 9-17 12.22 7-18

Mean SD Mean SD
LAPS 10.02 1.97 10.50 1.98
Income 6.15 2.26 6.08 2.26
Education 12.54 1.72 12.21 1.87
Occupation 602.90 355.90 241.20 141.60
Number of Children 2.19 .70 2.19 .70

 

Table 3. Means and standard deviations for mothers and fathers for each variable.

 

 

VARIABLE Mothers Fathers

Mean SD Mean SD
Percent Praise 4.00 4.30 2.00 2.20
Percent Command 35.00 13.10 37.00 15.00
Percent Critical 4.00 3.80 4.00 3.30
Percent Question 57.00 14.70 56.00 15.90
Hamilton Score 7.07 6.77 6.39 6.77
Beck Score 4.07 3.71 3.35 3.55

 

28

B k R l

Contrary to the hypotheses, ANCOVA analyses using Praise as the dependent
variable revealed no significant effects [F (1,163) = .56, NS]. ANCOVA analyses
using Critical Statements as the dependent variable also revealed no significant effects
[F (1,163) = 1.17, NS].

Mothers and fathers did use commands differently during the play interactions
with their sons. With Commands as the dependent variable, AN COVA analyses
revealed significant effects for both sex and depression level [F (1,163) = 9.8, p
<.0001]. Both sex (t = 2.19, p< .03) and Beck score (t = 3.97, p< .0001) affected the
number of commands given by the parents. For both mothers and fathers, the higher
the Beck score, the more commands were given, with mothers also giving significantly
more commands than fathers when depression was taken into account.

AN COVA analyses using Questions as the dependent variable also revealed
significant effects for both depression and sex [F(1,163) = 9.02, p< .0002]. The
higher the depression score (t = -3.86, p< .0002), the fewer questions were asked by
both mothers and fathers. Relative to the depression scores, fathers with higher
depression scores asked significantly fewer questions than mothers with higher

depression scores (t = -1.97, p< .05).

29
Hamilton ng Results
When Hamilton depression scores were used in ANCOVA analyses, results

were similar to Beck score results for Praise, i.e., Praise was not significant [F(1,163)
=.47, NS]. ANCOVA analyses using Critical Statements as the dependent variable
were also not significant when Hamilton scores were entered into the analyses
[F(1,163) = 2.0, NS).

AN COVA analyses using Commands as the dependent variable were significant
[F(1,163) = 3.58, p> .03). However, they differed from Beck results in that sex of the
parent (t = 2.045, p< .04) affected commands but Hamilton score (t.= 1.869, NS) did
not. Mothers did give significantly more commands than fathers, but depression did
not significantly affect the number of commands given by either sex when measmed
by Hamilton scores.

Finally, AN COVA analyses using Questions as the dependent variable revealed
significant effects [F(1,162) = 3.60, p> .03]. Hamilton Depression scores affected the
number of questions asked (t = -2.059,
p< .04), but sex had no effect (t = -1.864, NS) on the number of questions asked. The
higher the Hamilton score, the fewer questions were asked, regardless of sex of parent.

These results support the notion that depression does influence the way that
mothers speak to their children during play. However, fathers’ speech to their children
was also influenced by depression. This may reflect the instrumental nature of fathers’

interactions with children, which may be too strong to be offset by depression.

30
AN VA with ' 'onal variabl

In an attempt to explain what other variables might be accounting for variance
on the four dependent variables, (praise, commands, critical statements, and questions),
an AN COVA was run using several other variables. Lifetime Alcohol Problems
Scores (LAPS, Zucker, 1991), family income, parents’ education level, family
socioeconomic status (based on the Revised Duncan Socioeconomic Index, (Stevens &
Featlrerman, 1981), and the number of other children in the household were entered
into the ANCOVA analyses.

Results from these analyses were consistent with the above ANCOVA results.
With Commands as the dependent variable, AN COVA analyses that included the
additional variables mentioned above, were significant [F(1,162) = 3.73, p< .001].
However, the Beck Depression score accounted for the entire 14% of the variance
explained by these analyses (t = 2.376, p< .02). None of the other above mentioned
variables accounted for any significant portion of the variance in the number of
commands given.

The AN COVA analyses with Questions as the dependent variable, including
the additional variables, was also significant [F(1,162) = 3.55, p< .001], accounting for
14% of the variance. Similar to the Commands, Beck Depression Score accounted
for the entire 14% (t = -2.38, p< .02).

The AN COVA analyses for Critical statements and Praise with all additional

variables entered were not significant. [Praisez F(l,162) = .813, NS; Critical

31
statements: F(l,162) = 1.487, NS]. Again, this is consistent with the more specific

analyses above.

R ssion An s s

Stepwise multiple regression analyses were also conducted using all the above
variables (LAPS score, family income, parents’ education level, family occupational
status, and the number of other children in the household). This procedure was used
with the default significance level of .15 in order to identify whether any of the other
variables might be related to parental verbalizations in the playroom setting if
conditions were more stringent or sample sizes larger. Any of the variables that met
the . 15 level might be worthy of fm'ther investigation related to the Eyberg categories.

When Praise was entered as the dependent variable, only Family Income met
the . 15 significance level criteria for entry into the model. However, family income
did not significantly affect the amount of praise given [F(1,162) = 2.96, NS].

In the stepwise multiple regression using Commands as the dependent variable,
three other variables entered the model at the .15 significance level. Beck depression
score, parents’ education level, and family income were included in the analysis. Beck
Depression score accounted for the most variance (8%), [F(1,162) = 13.93, p< .0003],
again consistent with the AN COVA results above. Education level of the parents
accounted for a significant additional amount of variance (3%), [F(1,162) = 6.19, p<

.01] after Beck was accounted for. Income entered into the model at the .15 level, but

32

did not account for any significant additional amount of variance after Beck and
Education level were taken into account [F(1,162) = 2.83, NS].

Stepwise multiple regression using Critical Statements as the dependent
variable also allowed Family Income into the model at a significant level [F(1,162) =
4.63, p< .03], accounting for 2% of the variance. No other variables met the .15 level
for inclusion in the model for Critical statements.

When Questions was used as the dependent variable in stepwise multiple
regression, the three variables of Beck score, family income, and parents’ education
level again met the .15 significance level for entry. Beck Depression score accounted
for the most variance (8%), [F(1,162) = 13.29, p< .0004], followed by parents’
education level with an additional 10% ([F(1,162) = 5.73, p< .01]. Family income did
not account for any additional significant amount of variance [F(1,162) = 2.49, NS].

The regression analyses suggest that depression as measured by Beck is a
significant predictor of behavior in the playroom, based on the Eyberg coded
verbalizations. Parents’ education level also affecwd the Eyberg categories, and may
warrant further investigation as a predictor of parents’ behavior towards their children
in the play situation. With these analyses, none of the other variables (LAPS,

occupation, or number of children in family) accounted for additional variance.

Almhgl Problems Stags ES! Eyberg Qtegm’es

Finally, since the entire sample consisted of families with at least the father

being alcoholic, and since the mothers in the sample also tend to have higher alcohol

33

problems than control mothers, simple effects tests were done for each of the
dependent variables between the upper and lower quartiles of the sample. This was
done separately for mothers and fathers. These analyses were an attempt to
distinguish any differences between higher and lower levels of alcohol problems
within an all alcoholic sample.

For mothers, t-tests revealed a significant difference in LAPS scores between
the upper and lower quartiles [n = 42; upper quartile mean = 12.79, SD = 1.17; lower
quartile mean = 8.09, SD = .156; [t = 18.61, (1,41) p< .0001]. However, analysis of
covariance revealed no significant differences on any of the four dependent variables
of Praise, Critical statements, Commands, or Questions between the women in the
upper vs. the lower quartiles.

For males also, there was a significant difference between the upper and lower
quartile on LAPS score [n = 40; upper quartile mean = 13.15, SD = 1.17; lower
quartile mean = 8.186, SD = .80, t = 15.72 (1,39) p< .0001]. Quartile did not affect
any of the dependent variables, but Beck score was significantly related to Commands
(t = 2.95, p< .005) and to Questions (t = -2.87, p< .007). Consistent with the
AN COVA results reported above, higher depression scores on the Beck predicted more

commands and fewer questions for fathers.

DISCUSSION

Some of the specific hypotheses of this study were supported, others were not.
Based on Beck Depression Inventory scores, both more depressed mothers and more
depressed fathers gave more commands than less depressed mothers and fathers. This
is inconsistent with the hypothesis that depression would affect mothers’ commanding
behavior but not fathers’. However, the behavior of the mothers in this sample is
consistent with previous evidence that maternal depression was associated with verbal
punitiveness, particularly yelling and threatening (Susman et.al, 1985) because mothers
overall gave significantly more commands than fathers. Similarly, Longfellow et. a1
(1982) found severity of mothers’ depression was related to greater use of hostile and
controlling techniques, such as warnings and commands. The Longfellow et. al (1982)
study used a similar SES sample to the current study and also used observational data,
lending further support to the current findings.

Questions, viewed as a measure of engagement with the child, were expected
to be higher for fathers, but unrelated to depression in fathers. Contrary to this
hypothesis, both mothers and fathers asked significantly fewer questions as depression
scores increased. There was a slight sex difference in number of questions asked
when measmed by Beck only, with fathers actually asking fewer questions than
mothers as depression scores increased.

The prediction that fathers would be less likely to "take out" their depression

on their children was based on literature which views the father’s role as that of

34

35
playmate (Lamb, 1981; Yogman, 1982). Additionally, Patterson (1981) reported

observational data showing fathers have higher rates of play than mothers in both
distressed and non-distressed families. The current results suggest that in a sample of
alcoholic fathers, depression levels can affect the way fathers interact with their
children. Higher levels of depression in fathers, as well as mothers, was associated
with giving more commands to the child. Lower levels of depression in mothers were
expected to be related to higher rates of questions, and the findings support this.

The coding categories used in this study include two types of commands:
indirect and direct, and these may reflect different styles of interacting with the child.
According to the Eyberg Coding Manual, direct commands very clearly tell a child
what to do and always do so in a positive manner, i.e., they always tell a child what to
do, not what M to do. Indirect commands imply an expectation of the child but do
not use imperatives. For example, "Why don’t we play with the blocks?" Indirect
commands can be difficult to distinguish from questions, and the manual decision rule
reads: "uncertain between indirect command and question, code question." Numbers of
indirect vs. direct commands were not sufficient to divide them in the analyses. If
direct and indirect commands could be analyzed separately, it is possible that
depressed fathers would use more indirect commands and depressed mothers would
use more direct commands, indicating a more engaging style for fathers.

Results of the current study were similar for bath Beck and Hamilton measures
of depression in regard to number of questions asked during the play sessions; the

higher the depression score, the fewer questions were asked. It was predicted that sex

36

would also affect the number of questions asked, based on questions as a measure of
attempts to engage the child, and sex did affect Questions when measured by Beck
only but not at the significance level of the effect of the depression. It may be that in
lower socioeconomic level families, parents have an overall lower level of interaction
with their children, particularly as measured by attempts to engage them. Indeed, in
the current study, parents often seemed to engage in parallel play during the play
session rather than associative or cooperative play as one might expect in parental play
with preschool age children.

The possibility that parents in families characterized by alcoholism and
depression may use immature interactional styles during play activities with their
preschool age children can be tested in future studies by comparing high risk parents’
play style with that of comparison. Such comparison could be especially important in
revealing differences in the conduct or style of instrumental activities between
alcoholic and non-alcoholic fathers and their sons.

The results found in the questions category may reflect limitations of the
Eyberg measure. Following from the points made above, the questions category can
easily become a catch-all for many types of parental verbalizations. The coding
manual indicates that questions may be distinguished by inflection only, or when a tag
is added to a descriptive/reflective statement. This mode of 3 speech is very common
when speaking with young children. For example, "That’s a blue one, right " is coded
as a question, as is "I like it, don’t you?" The questions category tends to accumulate

the highest number of responses in any given protocol because of this catch-all nature.

37

It seems reasonable, therefore, to be cautious in interpreting the significant results in
this category. Since questions are usually coded more frequently than anything else in
a given protocol, and since the number of questions spoken decreased as the
depression scores increased (on both Hamilton and Beck), it may be that decreases in
frequency of questions is an indirect effect of a decrease in the overall level of
engagement with the child.

Other research using the same data set (Davies, Zucker, N 011, & Fitzgerald,
1991) as the current study (though not necessarily the same subjects) has shown levels
of paternal depression to be negatively correlated with positive affective parenting.
While this suggests that depressed fathers do not use positive parenting behaviors, the
current study goes one step farther and suggests that depression actually increases
negative parenting behavior for both mothers and fathers. 'Ihe Davies, et. a1 study
used factor analyzed self-report questionnaire data, the Child-Rearing Practices Report
(Block, 1980), to measure positive affect parenting. The current study used direct
observation of fathers’ behavior with similar results. More research is needed to
clarify the relationship between depression in fathers and interactions with their
children. Coyne & Downey (1990), in an recent review of the literature concerning
children of depressed parents, were unable to find a single published study in which
investigators observed depressed fathers interacting with their children.

The hypotheses regarding praise and critical statements were not supported. It
is quite likely that the Praise and Critical Statements analyses were not significant

because of the low rate of occurence of such statements. Eyberg criteria for coding

38

these categories are rigid, and were difficult to apply in the interactions with the
young children.

In the case of Critical statements, the Eyberg Coding Manual is perplexing.
Critical statements may be spoken in "sweet tones" but with a critical content.
Consider this interchange:

Child: "I want to build a big, big house."

Parent: "You can’t because there aren’t enough blocks."

Eyberg (Personal communication, June, 1989) codes this parent comment as a critical
statement. She argues that it "corrects" the child by pointing out "in nice tones" what
was wrong with the child’s idea. The parent response in this example could certainly
as easily be coded as a purely descriptive statement which has no negative reflection
on the child at all. Since the final coding note warns "if in doubt, DO NOT code
critical," it is likely that coders will err on the side of rendering controversial calls to
less severe categories.

Praise statements are coded as Labeled or Unlabeled Praise, depending on the
specificity. These two categories were combined for purposes of analysis in this study
because of their very low frequency of occurrence. According to the coding manual,
praise must reflect clear verbal approval, not just enthusiasm, and in the case of
Labeled Praise, it must be specific enough to let the child know exactly what he or she
would have to do to elicit the praise again. Praise statements may be in the form of
questions, such as "That picture turned out terrific, didn’t it?" However, as noted

above, tags on statements are extremely frequent when parents are talking to young

39
children, making the distinction between praise and questions difficult. Again, the

manual warns that "when in doubt, DO NOT code Praise." Some Praise statements
undoubtedly end up in the questions category.

Another problem with the Praise coding is that it may distort the "truth" of a
given situation. For example, if a parent says, "I like it when you sit still", Eyberg
codes praise (Personal communication, June, 1989). However, parental approval may
not always be equal to good. To follow the letter of the coding rules, if a parent said,
"I like it when you color pictures on the wallpaper" it would also be coded as Labeled
Praise.

These instrument drawbacks may explain the lack of significant results in the
Praise and Critical Statements categories. It is important to note that results were
exactly the same in these two categories, for both of the depression indices, as well as
for sex. Absolutely no relationships were found, even though previous research would
support the prediction that fathers would be more likely to praise their children than
mothers, and that perhaps depression in mothers would affect that behavior more than
depression in fathers. The converse would be true for critical statements if we think
of praise and critical statements as opposite ends of a continuum.

It is interesting that the women who are wives of alcoholics also report higher
levels of alcohol problems in relation to control group wives. Since in most cases
these women have not personally experienced as many alcohol related problems as the
men, e.g., arrests for drunk driving, physical aggression that led to fights, and do not

score nearly as high on measures of antisocial behavior (e. g., Fitzgerald, et. a1, under

40

review), it is possible that wives experience these problems as their own indirectly
through their husbands. The relationship between these alcohol problems and
depression and the way these are related to the way parents not only perceive their
childrens’ behavior but also how they actually speak to and interact with them is
fertile ground for further research using this data set. Other coding measures should be
used with the observational data, particularly instruments that are sensitive to affective
interactions between parents and their children.

Furthermore, another variable that was not examined in this study but that may
very well impact both depression and the types of interactions measured in this study
is antisociality.

Finally, if indeed it is supported with further research that parents actually
speak to their children differently when they are alcoholic or depressed or both, it
suggests interventions that may be effective in reducing the effects of depression on

child outcomes.

SUMMARY

This study lends support to the position that depression level affects the way
that parents speak to their children in a playtime situation. Sex differences were found
in two categories, supporting the notion that depression in mothers is likely to affect
children more than depression in fathers because depressed mothers use more negative
verbalizations overall than depressed fathers, assuming that this is balanced by the
amount of time spent with mothers vs. fathers as of course it was in this study.
According to this study, no evidence was found for fathers to use more positive

verbalizations, such as praise, as was predicted by previous research.

41

42

Table 1A. ANCOVA WITH SEX AND HAMILTON DEPRESSION ONLY
DEP VARIABLE: PRAI_PCT
ANALYSIS OF VARIANCE
SUM OF MEAN
SOURCE DF SQUARES SQUARE F VALUE PROB>F
MODEL 2 0.0014054% 0.000702703 0.474 0.6235
ERROR 162 0.24031199 0.001483407
C TOTAL 164 0.24171740
ROOT MSE 0.03851503 R-SQUARE 0.“)58
DEP MEAN 0.03088919 ADJ R-SQ 01sz
CV. 124.6877
PARAMETER ESTIMATES

NOTE: PARTIAL AND SEMI-PARTIAL CORRELATIONS ARE SQUARED

PARAMETER STANDARD T FOR H0:

VARIABLE DF ESTIMATE ERROR PARAMETER=0 PROB > rn
INTERCEP 1 0.03251540 0.(X)4231750 7.684 01an
SEX_DUM 1 -0.(X)2512718 0.(X)3(X)7017 -0.836 0.4046
HAMCURR l -0.(XX)251470 0010449684 -0.559 0.5768

SUM OF RESIDUALS 9.52276E-15

SUM OF SQUARED RESIDUALS 0.240312
PREDICI'ED RESID SS (PRESS) 024721183

43

Table 2A. ANCOVA WITH SEX AND HAMILTON DEPRESSION ONLY
DEP VARIABLE: COMM_PCT
ANALYSIS OF VARIANCE
SUM OF MEAN
SOURCE DF SQUARES SQUARE F VALUE PROB>F
MODEL 2 0.14483190 0.07241595 3.575 0.0302
ERROR 162 3.28 147991 0.02025605
C TOTAL 164 3.42631181
ROOT MSE 0.1423237 R-SQUARE 0.0423
DEP MEAN 0.3557933 ADJ R-SQ 0.0304
C.V. 40.m181
PARAMETER ESTIMATES

NOTE: PARTIAL AND SEMI-PARTIAL CORRELATTONS ARE SQUARED

PARAMETER STANDARD T FOR H0:

VARIABLE DF ESTIMATE ERROR PARAMETER=0 PROB > I'll
INTERCEP 1 0.335560% 0.01563749 21.459 0.1!!”
SEX_DUM l 0.02271983 0.01 l 11 176 2.045 0.0425
HAMCURR l 0.1!)3105383 0.(X)1661706 1.869 0.0635

SUM OF RESIDUALS 2.12164E-l3

SUM OF SQUARED RESIDUALS 3.28148

PREDICIED RESID SS (PRESS) 3.425593

44

Table 3A. ANCOVA WITH SEX AND HAMILTON DEPRESSION ONLY
DEP VARIABLE: CRIT_PCT
ANALYSIS OF VARIANCE

SUM OF MEAN
SOURCE DF SQUARES SQUARE F VALUE PROB>F
MODEL 2 0.005113040 0.“)2556520 1.994 0.1395
ERROR 162 0.20772555 0.(X)1282256
C TOTAL 164 0.21283859

ROOT MSE 0.0358086] R-SQUARE 0.0240
DEP MEAN 0.03751692 ADJ R-SQ 0.0120
C.V. 95.44656

PARAMETER ESTIMATES
NOTE: PARTIAL AND SEMI-PARTIAL CORRELATIONS ARE SQUARED

PARAMETER STANDARD T FOR H0:

VARIABLE DF ESTIMATE ERROR PARAMETER=0 PROB > I'll
INTERCEP l 0.03221608 01113934388 8.188 0.0001
SEX_DUM 1 0.001927890 0.(X)2795716 0.690 0.4914
HAMCURR 1 0.(XX)802573 0.m0418(185 1.920 0.0567

SUM OF RESIDUALS 1.36809E-14

SUM OF SQUARED RESIDUALS 0.w77256
PREDICIED RESID SS (PRESS) 0.2152816

45

Table 4A. ANCOVA WITH SEX AND HAMILTON DEPRESSION ONLY
DEP VARIABLE: QUES_PCI'
ANALYSIS OF VARIANCE
SUM OF MEAN
SOURCE DF SQUARES SQUARE F VALUE PROB>F
MODEL 2 0.16630698 0.08315349 3.594 0.0297
ERROR 162 3.74813246 0.02313662
C TOTAL 164 3.91443944
ROOT MSE 0.1521073 R-SQUARE 0.0425
DEP MEAN 0.5758006 ADJ R-SQ 0.0307
C.V. 26.41665
PARAMETER ESTIMATES

NOTE: PARTIAL AND SEMI-PARTIAL CORRELATIONS ARE SQUARED

PARAMETER STANDARD T FOR H0:

VARIABLE DF ESTIMATE ERROR PARAMETER=0 PROB > I'll
INTERCEP 1 0.59970828 0.01671243 35.884 01an
SEX_DUM 1 0.02213500 0.01187560 -1.864 0.0641
HAMCURR l 0.003656486 01111775934 -2.059 0.0411
SUM OF RESIDUALS 2.34576E-13

SUM OF SQUARED RESIDUALS 3.748132
PREDICTED RESID SS (PRESS) 3.914897

46

Table 5A. ANCOVA WITH SEX AND BECK DEPRESSION ONLY
DEP VARIABLE: PRAI_PCT
ANALYSIS OF VARIANCE
SUM OF MEAN
SOURCE DF SQUARES SQUARE F VALUE PROB>F
MODEL 2 0.001661526 0.(X)0830763 0.561 0.5720
ERROR 162 024005587 0.(X)l481826
C TOTAL 164 0.24171740
ROOT MSE 0.0384945 R-SQUARE 0.(X)69
DEP MEAN 0.03088919 ADJ R-SQ -0.(X)54
C.V. 124.6213
PARAMETER ESTIMATES

NOTE: PARTIAL AND SEMI-PARTIAL CORRELATIONS ARE SQUARED

PARAMETER STANDARD T FOR H0:

VARIABLE DF ESTIMATE ERROR PARAMETER=0 PROB > I'll
INTERCEP 1 0.03294406 0.1!)4247960 7.755 0.00m
SEX_DUM 1 «0.002498015 0.003001354 -0.832 0.4065

BECK 1 -0.0(X)577581 0.(XX)828592 -0.697 0.4868

SUM OF RESIDUALS 9.59389E-15

SUM OF SQUARED RESIDUALS 0.2400559
PREDICIED RESID SS (PRESS) 0.248505

47

Table 6A. ANCOVA WITH SEX AND BECK DEPRESSION ONLY
DEP VARIABLE: COMM_PCT
ANALYSIS OF VARIANCE
SUM OF MEAN
SOURCE DF SQUARES SQUARE F VALUE PROB>F
MODEL 2 0.37089150 0.18544575 9.832 0.0(I)I
ERROR 162 3.05542031 0.01886062
C TOTAL 164 3.42631181
ROOT MSE 0.137334 R-SQUARE 0.1082
DEP MEAN 0.3557933 ADJ R-SQ 0.0972
C.V. 38.59937
PARAMETER ESTIMATES

NOTE: PARTIAL AND SEMI-PARTIAL CORRELATIONS ARE SQUARED

PARAMETER STANDARD T FOR H0:

VARIABLE DF ESTIMATE ERROR PARAMETER=0 PROB > [11
INTERCEP 1 0.31357632 0.01515513 20.691 0.0(XJ1
SEX_DUM 1 0.02340422 0.01070771 2.186 0.0303
BECK l 0.01172668 01112956105 3.967 0.0(XJ1

SUM OF RESIDUALS 2.11706E-13

SUM OF SQUARED RESIDUALS 3.05542
PREDICI'ED RESID SS (PRESS) 3.174559

48

Table 7A. ANCOVA WITH SEX AND BECK DEPRESSION ONLY
DEP VARIABLE: CRIT_PCI'
ANALYSIS OF VARIANCE
SUM OF MEAN
SOURCE DF SQUARES SQUARE F VALUE PROB>F
MODEL 2 0.003018406 01111509203 1.165
ERROR 162 0.20982019 0.1!)1295186
C TOTAL 164 0.21283859
ROOT MSE 0.0359887 R-SQUARE 0.0142
DEP MEAN 0.03751692 ADJ R-SQ 0.11120
C.V. 95.92658
PARAMETER ESTIMATES

NOTE: PARTIAL AND SEMI-PARTIAL CORRELATIONS ARE SQUARED

PARAMETER STANDARD TFOR H0:

VARIABLE DF ESTIMATE ERROR PARAMETER=0 PROB > lTl
INTERCEP 1 0.03353410 01113971439 8.444 MIDI
SEX_DUM l 0.(X)1741596 0.“)2805981 0.621 0.5357
BECK 1 0.001103983 0.(XX)774655 1.425 0.1560
SUM OF RESIDUALS 1.35014E-14

SUM OF SQUARED RESIDUALS 0.w98202

PREDICIED RESID SS (PRESS) 0.2193368

49

Table 8A. ANCOVA WITH SEX AND BECK DEPRESSION ONLY
DEP VARIABLE: QUES_PCI'
ANALYSIS OF VARIANCE
SUM OF MEAN
SOURCE DF SQUARES SQUARE F VALUE PROB>F
MODEL 2 0.39227439 0.19613719 9.021 0.0002
ERROR 162 3.52216506 0.02174176
C TOTAL 164 3.91443944
ROOT MSE 0.1474509 R-SQUARE 0.1002
DEP MEAN 0.5758006 ADJ R-SQ 0.0891
C.V. 25.60797
PARAMETER ESTIMATES

NOTE: PARTIAL AND SEMI-PARTTAL CORRELATIONS ARE SQUARED

PARAMETER STANDARD T FOR H0:

VARIABLE DF ESTIMATE ERROR PARAMETER=0 PROB > m
INTERCEP 1 0.61994552 0.01627156 38.100 0.0001
SEX_DUM 1 0.02264780 0.01 149651 -1 .970 0.0505
BECK 1 0.01225308 0.(X)3173871 -3.861 0.11112
SUM OF RESIDUALS 2.33660E-13

SUM OF SQUARED RESIDUALS 3.522165
PREDICI'ED RESID SS (PRESS) 3.654302

50

Table 9A. ANCOVA WITH ALL VARIABLES ENTERED

DEP VARIABLE: PRAI_PCI'

ANALYSIS OF VARIANCE
SUM OF MEAN
SOURCE DF SQUARES SQUARE F VALUE PROB>F
MODES 7 0.008503187 0.(X)1214741 0.813 0.5792
ERROR 156 0.23309930 0.(X)1494226
C TOTAL 163 0.24160249

ROOT MSE 0.03865522 R-SQUARE 0.0352
DEP MEAN 0.03095435 ADJ R—SQ -0.(X)81
C.V. 124.8782

PARAMETER STANDARD T FOR H0:

VARIABLE DF ESTIMATE ERROR PARAMETER=O PROB > ITI
INTERCEP 1 0.01665602 0.03231094 0.515 0.659
LAPS 1 0.0m240064 01111662743 0.144 0.8854
INCOME 1 0.(X)2222709 0.(X11540917 1.442 0.1512
EDUCA 1 0.000127022 0.1!)1859738 0.068 0.9456
OCCUP 1 0.000014591 .w000970892 1.503 0.1349
KIDS 1 -0.003488426 0.004399110 -0.793 0.4290
HAMCURR 1 —0.0(X)150425 0.(XJ0471542 -0.319 0.7501
BECK 1 ~0.(XX)247382 0.(XX)912908 0.271 0.7868
SUM OF RESIDUALS 9.42562E-15

SUM OF SQUARED RESIDUALS 0.2330993
PREDICIED RESID SS (PRESS) 0.254302

51

Table 10A. ANCOVA WITH ALL VARIABLES ENTERED

DEP VARIABLE: COMM_PCT

ANALYSIS OF VARIANCE
SUM OF MEAN

SOURCE DF SQUARES SQUARE F VALUE PROB>F
MODEL 7 0.4803162] 015861660 3.730 0.0010
ERROR 156 2.86991768 0.01839691
C TOTAL 163 3.35023389

ROOT MSE 0.1356352 R-SQUARE 0.1434

DEP MEAN 0.35747 ADJ R-SQ 0.1049

C.V. 37.9431

PARAMETER STANDARD T FOR H0:

VARIABLE DF ESTIMATE ERROR PARAMETER=0 PROB > ITI
INTERCEP 1 0.43713626 0.11337411 3.856 0.(XX)2
LAPS 1 0.005872849 0.“)5834307 1.(X)7 0.3157
INCOME 1 -0.007521131 01115406839 -1.391 0.1662
EDUCA 1 -0.01125986 0.(X)6525533 -1.726 0.0864
OCCUP 1 0.(XX)041077 0.(X)0034(E7 1.206 0.2297
KIDS 1 0.002151413 0.01543580 -0.139 0.8893
HAMCURR l 0.0(D833133 01111654569 0.504 0.6153
BECK 1 0.007611537 0.1XJ3203252 2.376 0.0187
SUM OF RESIDUALS 2.02241E-13

SUM OF SQUARED RESIDUALS 2.869918
PREDICTED RESID SS (PRESS) 3.179577

52

Table 11A. ANCOVA WITH ALL VARIABLES ENTERED

DEP VARIABLE: CRIT_PCT

ANALYSIS OF VARIANCE
SUM OF MEAN

SOURCE DF SQUARES SQUARE F VALUE PROB>F
MODEL 7 0.01322644 0.“)1889491 1.487 0.1746
ERROR 156 0.19819605 0.“)1270488
C TOTAL 163 0.21142249

ROOT MSE 0.0356439 R-SQUARE 0.0626

DEP MEAN 0.03774568 ADJ R-SQ 0.0205

C.V. 94.43173

PARAMETER STANDARD T FOR H0:

VARIABLE DF ESTIMATE ERROR PARAMETER=0 PROB > ITI
INTERCEP 1 0.08860604 0.02979385 2.974 0.0034
LAPS 1 -0.(X)2283175 0.(I)153321 1 -1.489 0.1385
INCOME l -0.001898576 0.(X)l420876 -1.336 0.1834
EDUCA 1 -0.(X)l 186805 01111714860 0.692 0.4899
OCCUP 1 -.0(X)0087769 111000895257 -0.980 0.3284
KIDS 1 0002249725 0.(X)4056410 -0.555 0.581!)
HAMCURR l 0.(XX)669626 01110434808 1.540 0.1756
BECK 1 0.(XX)851551 01110841790 1.012 0.3133

SUM OF RESIDUALS 3.1!)046E-14

SUM OF SQUARED RESIDUALS 0.1981961
PREDICTED RESID SS (PRESS) 0.2222447

53

PROB > ITI

Table 12A. ANCOVA WITH ALL VARIABLES ENTERED
DEP VARIABLE: QUES_PCI‘
ANALYSIS OF VARIANCE
SUM OF MEAN
SOURCE DF SQUARES SQUARE F VALUE PROB>F
MODEL 7 0.52406033 0.07486576 3.555
ERROR 156 3.28529092 0.02105956
C TOTAL 163 3.80935125
ROOT MSE 0.1451191 R-SQUARE 0.1376
DEP MEAN 0.57383 ADJ R-SQ 0.0989
C.V. 25.28957
PARAMETER STANDARD T FOR H0:
VARIABLE DF ESTIMATE ERROR PARAMETER=0
INTERCEP l 0.45760168 0.12130148 3.772 0.01112
LAPS l 0.003829739 0.“)6242254 -0.614 0.5404
INCOME 1 0.007 196999 0.(X)5784897 1.244 0.2153
EDUCA 1 0.01231964 0.(X)6981813 1.765 0.0796
OCCUP 1 -0.0(X)046892 0.“)0036449 -1.287 0.21112
KIDS 1 0.007889564 0.01651510 0.478 0.6335
HAMCURR 1 0001352334 0.1!)1770260 -0.764 0.4461
BECK 1 -0.008215706 0.003427231 -2.397 0.0177
SUM OF RESIDUALS 2.60653E-13
SUM OF SQUARED RESIDUALS 3.285291
PREDICI'ED RESID SS (PRESS) 3.646666

5O
45
4O
35
80
25
20
15
10

Figure 1A.

54

Percent Commands

 

I

l

I

 

l l J l l 1 l l J l

 

 

2 8 4 5 6 7 8 9 10 11
Beck Depression

—— Females -4— Males

Relationship between Beck Depression and Percent Commands by Sex

12

13

Percent Questions

55

 

80

70"

l

60
5O

BOr

20”

l

10

 

l l l l l J L l

 

 

5 e 7 8 9 10 11 12
Beck Depression

-*- Females ‘4’ Maies

Figure 2A. Relationship Betweeen Beck Depression and Percent Questions by Sex

18

56

Percent Commands

 

80

60r

 

 

 

80-

I

10

 

 

 

O l 1 l l l l l l 1 l l l l 1 l l 1 1

12 8 4 5 6 7 8 91011121814151617181920
Hamilton DepreSSlon

-‘- Females —*- Males

Figure 3A. Relationship Between Hamilton Depression and Percent Commands by Sex

57

Percent Questions

 

80
70 r

60 L a-

l

50 1
40 P
30 —
20 b

10”

 

 

l l l l 1 l 1 l l 1 l l l

1 l L 1 l
1284567891011121814151617181920
Hamilton Depression

 

O

—'-- Females —4— Males

Figure 4A. Relationship Between Hamilton Depression and Percent Questions by Sex

58

Table 13A. STEPWISE MULTIPLE REGRESSION WITH ALL VARIABLES ENTERED

STEPWISE REGRESSION PROCEDURE FOR DEPENDENT VARIABLE PRAI_PCI'

NOTE: SLENTRY AND SLSTAY HAVE BEEN SET TO
.15 FOR THE STEPWISE TECHNIQUE.

STEP 1 VARIABLE INCOME ENTERED R SQUARE = 0.01793227

C(P) = -1.20878575
DF SUM OF MEAN SQUARE F PROB>F
SQUARES
REGRESSION 1 0.111433248 0111433248 2.96 0.0874
ERROR 162 0.23727001 0.ml46463
TOTAL 163 0.24160249
B VALUE STD ERROR TYPE 11 SS F PROB>F
INTERCEPT 0.01682358
INCOME 0.(X)229223 0.(X)133277 0.m433248 2.96 0.0874

 

NO OTHER VARIABLES MET THE 0.1500 SIGNIFICANCE LEVEL FOR ENTRY

 

SUMMARY OF STEPWISE REGRESSION PROCEDURE FOR DEPENDENT VARIABLE

PRAI_PCT
VARIABLE NUMBER PARTIAL MODEL
STEP ENTERED REMOVED IN R"2 R“2 C(P)
1 INCOME 1 0.0179 0.0179 -1.20879
VARIABLE
STEP ENTERED REMOVED F PROB>F

1 INCOME 2.9581 ' 0.0874

59
Table 14A. STEPWISE MULTIPLE REGRESSION WITH ALL VARIABLES ENTERED

STEPWISE REGRESSION PROCEDURE FOR DEPENDENT VARIABLE COMM_PCT

NOTE: SLENTRY AND SLSTAY HAVE BEEN SET TO
.15 FOR TI-E STEPWISE TECHNIQUE.

STEP 1 VARIABLE BECK ENTERED R SQUARE = 0.07918045
C(P) = 7.68909337
DF SUM OF MEAN SQUARE F PROB>F
SQUARES
REGRESSION 1 0.26527303 0.26527303 13.93 0.0(XJ3
ERROR 162 3.08496086 0.01904297
TOTAL 163 3.35023389

B VALUE STD ERROR TYPE 11 SS F PROB>F

INTERCEPT 0.31696715

 

BECK 0.01108926 0.00297114 026527303 13.93 0.(X)03
BOUNDS ON CONDITION NUMBER: 1, 1
STEP 2 VARIABLE EDUCA ENTERED R SQUARE =- 0.11328179
C(P) = 3.47894920
DF SUM OF MEAN SQUARE F PROB>F

SQUARES
REGRESSION 2 0.37952048 0.18976024 10.28 0.11101
ERROR 161 2.97071341 0.01845164
TOTAL 163 3.35023389

B VALUE STD ERROR TYPE II SS F PROB>F
INTERCEPI' 0.50458475
EDUCA -0.01492767 010599910 0.11424745 6.19 0.0139
BECK 0.01031 1 10 0.“)294 132 0.22675703 12.29 0.0M6

BOUNDS ON CONDITION NUMBER: 1.011433, 4.045734

 

STEP 3 VARIABLE INCOME ENTERED R SQUARE = 0.12869616

C(P) = 2.67186012
DF SUM OF SQUARES MEAN SQUARE F PROB>F
REGRESSION 3 0.43116224 0.14372075 7.88 01an
ERROR 160 2.91907165 0.01824420

TOTAL 163 3.35023389

B VALUE
INTERCEPT 0.51746955
INCOME -0.(X)887369
EDUCA -0.01 1 12153
BECK 0.“)886156

60

STD ERROR TYPE 11 SS F PROB>F
0111527432 0.05164176 2.83 0.0944
0111637985 0.055441 11 3.04 0.0832
01113049“) 0.15410967 8.45 0.0042

NO OTIER VARIABLES MET TIE 0.1500 SIGNIFICANCE LEVEL FOR ENTRY

 

SUMMARY OF STEPWISE REGRESSION PROCEDURE FOR DEPENDENT VARIABLE

COMM_PCT
VARIABLE NUMBER PARTTAL MODEL
STEP ENTERED REMOVED IN R"2 R“2 C(P)
1 BECK 1 0.0792 0.0792 7.68909
2 EDUCA 2 0.0341 0.1133 3.47895
3 INCOME 3 0.0154 0.1287 2.67186
VARIABLE

STEP ENTERED REMOVED F PROB>F

1 BECK 13.9302 0.01113

2 EDUCA 6.1917 0.0139

3 INCOME 2.8306 0.0944

61

Table 15A. STEPWISE MULT'IPLE REGRESSION WITH ALL VARIABLES ENTERED

STEPWISE REGRESSION PROCEDURE FOR DEPENDENT VARIABLE CRIT_PCI‘

NOTE: SLENTRY AND SLSTAY HAVE BEEN SET TO
.15 FOR TIE STEPWISE TECHNIQUE.

STEP 1 VARIABLE INCOME ENTERED R SQUARE = 0.02780965

C(P) = 1.78270135
DF SUM OF MEAN SQUARE F PROB>F
SQUARES
REGRESSION 1 0.(X)587959 0.“)587959 4.63 0.0328
ERROR 162 0.20554290 0.(X)126878
TOTAL 163 0.21 142249
B VALUE STD ERROR TYPE 11 SS F PROB>F
INTERCEPT 0.05420724
INCOME 0111267032 0.1]1124046 0.(X)587959 4.63 0.0328

 

NO OTHER VARIABLES MET TIE 0.1500 SIGNIFTCANCE LEVEL FOR ENTRY

 

SUMMARY OF
STEPWISE REGRESSION PROCEDURE FOR DEPENDENT VARIABLE CRIT_PCI'
VARIABLE NUMBER PARTTAL MODEL
STEP ENTERED REMOVED IN R"2 R"2 C(P)
l INCOME 1 0.0278 0.0278 1.78270
VARIABLE
STEP ENTERED REMOVED F PROB>F

1 INCOME 4.6340 0.0328

62

Table 16A.

STEPWISE MULTIPLE REGRESSION WITH ALL VARIABLES ENTERED

STEPWISE REGRESSION PROCEDURE FOR DEPENDENT VARIABLE QUES_PCT

NOTE: SLENTRY AND SLSTAY HAVE BEEN SET TO
.15 FOR TIE STEPWISE TECHNIQUE.

 

 

STEP 1 VARIABLE BECK ENTERED R SQUARE = 0.07579291
C(P) = 7.17490383
DF SUM OF MEAN SQUARE F PROB>F
SQUARES
REGRESSION 1 0.28872180 0.28872180 13.29 0.01114
ERROR 162 3.52062945 0.02173228
TOTAL 163 3.80935125
B VALUE STD ERROR TYPE 11 SS F PROB>F
INTERCEPT 0.61608503
BECK -0.01 1569“) 0.00317401 0.28872180 13.29 0.0004
STEP 2 VARIABLE EDUCA ENTERED R SQUARE = 0.10754361
C(P) = 3.43168773
DF SUM OF MEAN SQUARE F PROB>F
SQUARES
REGRESSION 2 0.40967139 0.m483569 9.70 0.01111
ERROR 161 3.39967986 0.0211 1602
TOTAL 163 3.80935125
B VALUE STD ERROR TYPE 11 SS F
PROB>F
INTERCEPT 0.42304273
EDUCA 0.01535928 0.(X)641763 0.12094959 5.73 0.0178
BECK -0.01076834 0.(X)314652 0.24731377 1 1.71 011108
STEP 3 VARIABLE INCOME ENTERED R SQUARE = 0.12121527
C(P) = 2.95869470
DF SUM OF MEAN SQUARE F PROB>F
SQUARES
REGRESSION 3 0.46175153 0.15391718 7.36 0.0(XJ2
ERROR 160 3.34759972 0.02092250
TOTAL 163 3.80935125

63

B VALUE STD ERROR TYPE 11 SS F PROB>F
INTERCEPT 0.41010335
INCOME 0.(X)891 128 0.(X)564821 0.05208014 2.49 0.1 166
EDUCA 0.01153702 0.(X)683211 0.05966098 2.85 0.0932
BECK -0.(X)931266 0.1!)326514 0.17019903 8.13 0.0049N

NO OTIER VARIABLES MET TIE 0.1500 SIGNIFICANCE LEVEL FOR ENTRY

 

SUMMARY OF
STEPWISE REGRESSION PROCEDURE FOR DEPENDENT VARIABLE QUES_PCT
VARIABLE NUMBER PARTTAL MODEL
STEP ENTERED REMOVED IN R"2 R“2 C(P)
1 BECK 1 0.0758 0.0758 7.17490
2 EDUCA 2 0.0318 0.1075 3.43169
3 INCOME 3 0.0137 0.1212 2.95869
VARIABLE
STEP ENTERED REMOVED F PROB>F
1 BECK 13.2854 011304
2 EDUCA 5.7279 0.0178
3 INCOME 2.4892 0.1 166

Table 17A. MALE ONLY T-TEST: T-TEST PROCEDURE

VARIABLE: LAPS

QUART N MEAN STD DEV STD ERROR
1 20 13.15410000 1.16558708 0.26063320
-1 20 8.18615000 0.79965461 0.17880821

VARIANCES T DF PROB > ITI

UNEQUAL 15.7177 33.6 0.0001
EQUAL 15.7177 38.0 0.0001

FOR H0: VARIANCES ARE EQUAL, F: 2.12 WITH 19 AND 19 DF
PROB > F’= 0.1090

 

Table 18A. HAMILTON RESULTS

MALE ONLY r-rssr
DEP VARIABLE: PRALPCT _
ANALYSIS OF VARIANCE
SUM OF MEAN
SOURCE OF SQUARES SQUARE F VALUE PROB>F
MODEL 2 0000794709 0000397355 1.202 0.3118
ERROR 38 0.01256368 0.000330623

C TOTAL 40 0.01335839
ROOT MSE 0.01818305 R-SQUARE 0.0595
DEP MEAN 0.01443597 ADJ R-SQ 0.0100
C.V. 125.9566

PARAMETER ESTTMATES

PARAMETER STANDARD T FOR H0:

VARIABLE DF ESTIMATE ERROR PARAMETER=0 PROB > ITI
INTERCEP 1 0.01550356 0.(X)3839328 4.038 0.0003
QUART 1 0.004684408 01133045663 1.538 0.1323

HAMCURR 1 -0.0(X)145304 01110389600 -0.373 0.7113

 

65
Table 19A. MALE ONLY T-TEST

DEP VARIABLE: COMM_PCT

 

ANALYSIS OF VARIANCE
SUM OF MEAN
SOURCE DF SQUARES SQUARE F VALUE PROB>F
MODEL 2 0.05348160 0.02674080 0.871 0.4268
ERROR 38 1.16679105 0.03070503
C TOTAL 40 1.22027265
ROOT MSE 0.1752285 R-SQUARE 0.0438
DEP MEAN 0.3881033 ADJ R—SQ 41.11165
C.V. 45.14996
PARAMETER ESTIMATES
PARAMETER STANDARD T FOR H0:
VARIABIE DF ESTIMATE ERROR PARAMETER=0 PROB > ITI
INTERCEP 1 0.39316354 0.03699928 10.626 0.01111
QUART 1 0.03761390 0.02935080 1.282 0.2078
HAMCURR l -0.0(X)631436 01113754543 -0.168 0.8673
Table 20A. MALE ONLY T-TEST
DEP VARIABIE: CRIT_PCI'
ANALYSIS OF VARIANCE
SUM OF MEAN
SOURCE DF SQUARES SQUARE F VALUE PROB>F
MODEL 2 0.004963235 0102481618 2.146 0.1309
ERROR 38 0.04393794 0.(I)1 156262
C TOTAL 40 0.04890118
ROOT MSE 0.03400385 R-SQUARE 0.1015
DEP MEAN 0.04676715 ADJ R-SQ 0.0542
C.V. 72.71382
PARAMETER ESTIMATES
PARAMETER STANDARD T FOR H0:
VARIABLE DF ESTIMATE ERROR PARAMETER=0 PROB > 1T1
INTERCEP l 0.04275927 01117 179870 5.955 0.(XX)1
QUART 1 -0.01179892 01115695650 -2.072 0.0451

HAMCURR 1 0.0(XJ5670M 01110728585 0.778

0.4413

 

Table 21A. MALE ONLY T-TEST

DEP VARIABLE: QUES_PCT

 

ANALYSIS OF VARIANCE
SUM OF MEAN
SOURCE DF SQUARES SQUARE F VALUE PROB>F
MODEL 2 0.03674775 0.01837388 0.558 0.5769
ERROR 38 1.25083468 0.03291670
C TOTAL 40 128758243
ROOT MSE 0.1814296 R-SQUARE 0.0285
DEP MEAN 0.5506936 ADJ R-SQ -0.0226
C.V. 32.94565
PARAMETER ESTIMATES
PARAMETER STANDARD T FOR H0:
VARIABLE DF ESTIMATE ERROR PARAMETER=0 PROB > ITI
INTERCEP 1 0.54857363 0.03830864 14.320 0.0001
QUART 1 0.03049939 0.03038949 -1.(X)4 0.3219
HAMCURR 1 0.0(13209734 0.(X)3887411 0.054 0.9573
Table 22A. BECK DEPRESSION RESULTS
MALE ONLY T-IEST
DEP VARIABIE: PRAI_PCT
ANALYSIS OF VARIANCE
SUM OF MEAN
SOURCE DF SQUARES SQUARE F VALUE PROB>F
MODEL 2 0.0m755634 01110377817 1.139 0.3308
ERROR 38 0.01260276 0.1110331652
C TOTAL 40 0.01335839
ROOT MSE 0.0182113 R-SQUARE 0.0566
DEP MEAN 0.01443597 ADJ R-SQ 0.11169
C.V. 126.1523
PARAMETER ESTIMATES
PARAMETER STANDARD T FOR H0:
VARIABLE DF ESTIMATE ERROR PARAMETER=0 PROB > ITI
INTERCEP 1 0.01492450 0.“)3895935 3.831 0.01115
QUART 1 0.1X14447663 0.“)3087150 1.441 0.1579
BECK 1 -0.0(X)113740 0.(I)0787806 -0.144 0.8860

 

67

 

Table 23A. MALE ONLY T-TEST
DEP VARIABIE: COMM_PCT
ANALYSIS OF VARIANCE
SUM OF MEAN
SOURCE DF SQUARES SQUARE F VALUE PROB>F
MODEL 2 0.27072073 0.13536037 5.417 0.0085
ERROR 38 0.94955192 0.02498821
C TOTAL 40 122027265
ROOT MSE 0.1580766 R-SQUARE 0.2219
DEP MEAN 0.3881033 ADJ R-SQ 0.1809
C.V. 40.73055
PARAMETER ESTIMATES
PARAMETER STANDARD T FOR H0:
VARIABLE DF ESTIMATE ERROR PARAMETER=0 PROB > ITI
INTERCEP 1 0.32072032 0.03381725 9.484 0.0(Xll
QUART 1 0.005095824 002679689 0.190 0.8502
BECK 1 0.02020290 01116838265 2.954 0.0054
Table 24A. MALE ONLY T-IEST
DEP VARIABLE: CRIT_PCI'
ANALYSIS OF VARIANCE
SUM OF MEAN
SOURCE DF SQUARES SQUARE F VALUE PROB>F
MODEL 2 0.004336759 0.1112168379 1.849 0.1713
ERROR 38 0.04456442 0.1111 172748
C TOTAL 40 0.04890118
ROOT MSE 0.03424541 R-SQUARE 0.0887
DEP MEAN 0.04676715 ADJ R-SQ 0.0407
C.V. 73.22534
PARAMETER ESTIMATES
PARAMETER STANDARD T FOR H0:
VARIABLE DF ESTIMATE ERROR PARAMETER=0 PROB > ITI
INTERCEP 1 0.04526280 01117326102 6.178 0.0(XJI
QUART 1 -0.01076523 0.m5805225 -1.854 0.0715
BECK 1 0.0111371629 0.(X)1481429 0.251 0.8033

 

68

Table 25A. MALE ONLY T-TEST

DEP VARIABLE: QUES_PCT

ANALYSIS OF VARIANCE
SUM OF MEAN
SOURCE DF SQUARES SQUARE F VALUE PROB>F
MODEL 2 0.26036335 0.13018167 4.816 0.0137
ERROR 38 1.027219% 0.027032%
C TOTAL 40 128758243
ROOT MSE 0.1644144 R-SQUARE 0.2022
DEP MEAN 0.5506936 ADJ R-SQ 0.1602
C.V. 29.85587
PARAMETER ESTIMATES
PARAMETER STANDARD T FOR H0:
VARIABIE DF ESTIMATE ERROR PARAMETER-=0 PROB > ITI
INTERCEP 1 0.61909237 0.035173% 17.601 0.0%1
QUART 1 0.001221746 0.02787125 0.044 0.9653

BECK l 0W9 00171 12432 -2.877 0.0%6

 

 

69

Table 26A. FEMALE ONLY T—TEST

T-TEST PROCEDURE
VARIABLE: LAPS
QUART N MEAN STD DEV STD ERROR
1 20 12.78855000 1.17357360 0.262419%
-1 22 8.09122727 0.15663301 0.03339427

VARIANCES T DF PROB > ITI

UNEQUAL 17.7569 19.6 0.0001
EQUAL 18.6149 40.0 0.1XXJI
FOR H0: VARIANCES ARE EQUAL, F: 56.14 WITH 19 AND 21 DF

PROB > F’= 0.0001

 

Table 27A. FEMALE ONLY T—TEST
HAMILTON DEPRESSION RESULTS

DEP VARIABLE: PRAI_PCT

ANALYSIS OF VARIANCE
SUM OF MEAN
SOURCE DF SQUARES SQUARE F VALUE PROB>F
MODEL 2 0.004576533 00122882167 1.731 0.1901
ERROR 40 0.05287041 0.%1321760
C TOTAL 42 0.05744694

ROOT MSE 0.036356% R-SQUARE 0.0797
DEP MEAN 0.03907227 ADJ R-SQ 0.0336
C.V. 93.04814

PARAMETER ESTIMATES
PARAMETER STANDARD T FOR H0:
VARIABLE DF ESTIMATE ERROR PARAMETER=0
INTERCEP 1 0.0481 1 110 0.%8428407 5.7%

QUART 1 0.%7771042 0015578158 1.393
HAMCURR 1 -0.(XX)992817 0.(XX)736503 -1.348

PROB) ITI

0.(XX)1
0.1713
0.1852

 

70

Table 28A. FEMALE ONLY T-TEST

DEP VARIABIE: COMM.PCI‘

 

ANALYSIS OF VARIANCE
SUM OF MEAN
SOURCE DF SQUARES SQUARE F VALUE PROB>F
MODEL 2 0.01222367 0.(X)61 1 1837 0.386 0.6825
ERROR 40 0.63385713 0.01584643
C TOTAL 42 0.646%%0
ROOT MSE 0.1258826 R-SQUARE 0.0189
DEP MEAN 0.3534683 ADJ R-SQ ~0.0301
C.V. 35.61354
PARAMETER ESTIMATES
PARAMETER STANDARD T FOR H0:
VARIABLE DF ESTIMATE ERROR PARAMETER-110 PROB > ITI
INTERCEP 1 0.3509934] 012918333 12.027 0.0%]
QUART l 0.01632533 0.01931435 0.845 0.4030
HAMCURR 1 0.0%422274 01112550140 0.166 0.8693
Table 29A. FEMALE ONLY T—TEST
DEP VARIABLE: CRIT_PCI'
ANALYSIS OF VARIANCE
SUM OF MEAN
SOURCE DF SQUARES SQUARE F VALUE PROB>F
MODEL 2 0.%5265151 01112632576 1.330 0.2759
ERROR 40 0.07917358 01111979340
C TOTAL 42 0.%443873
ROOT MSE 0.0444897? R-SQUARE 01524
DEP MEAN 0.03562748 ADJ R-SQ 0.0155
C.V. 124.8749
PARAMETER ESTIMATES
PARAMETER STANDARD T FOR H0:
VARIABLE DF ESTIMATE ERROR PARAMETER=0 PROB > IT]
INTERCEP 1 0.02609383 0.010314% 2.530 0.0154
QUART 1 0.1X50303m 0.%6826131 0.883 0.3823

HAMCURR 1 0%] 163147 0.%0901278 1.291

0.2043

 

71

Table 30A. FEMALE ONLY T-TEST
DEP VARIABLE: QUES_PCT

 

ANALYSIS OF VARIANCE
SUM OF MEAN
SOURCE DF SQUARES SQUARE F VALUE PROB>F
MODEL 2 0.04068770 0.02034385 0.973 0.3869
ERROR 4o 0.83676267 0.02091907
c TOTAL 42 0.87745037 )
ROOT MSE 0.1446343 R—SQUARE 0.0464
DEP MEAN 0.5718319 ADJ R-SQ 0.0013
C.V. 25.29314
PARAMETER ESTIMATES
PARAMETER STANDARD T FOR H0:
VARIABLE DF ESTIMATE ERROR PARAMETER=0 PROB > rn
INTERCEP 1 0.57480165 003353052 17.143 0.0001
QUART 1 0.03012667 0.02219145 -1.358 0.1822
HAMCURR 1 0.000592604 0002930013 0202 0.8407

 

Table 31A. FEMALE ONLY T-TEST

BECK DEPRESSION RESULTS
DEP VARIABLE: PRAI_PCT
ANALYSIS OF VARIANCE
SUM OF MEAN
SOURCE DF SQUARES SQUARE F VALUE PROB>F
MODEL 2 0.%3781 148 0.%1890574 1.409 0.2562
ERROR 40 0.05366579 0.%134]645

C TOTAL 42 0.05744694

ROOT MSE 0.03662847 R-SQUARE 0.%58
DEP MEAN 0.03907227 ADJ R-SQ 0.0191

C.V. 93.74544
PARAMETER ESTIMATES
PARAMETER STANDARD T FOR H0:
VARIABLE DF ESTIMATE ERROR PARAMETER=0 PROB > IT]
INTERCEP 1 004792842 0.%9471324 5.060 01an
QUART 1 0.1XJ9483899 0.%5998775 1.581 0.1218

BECK I -0.%15872% 0.%1450518 -1.094 0.2804

 

72
Table 32A. FEMALE ONLY T-TEST

DEP VARIABLE: COMM_PCT

 

ANALYSIS OF VARIANCE
SUM OF MEAN
SOURCE DF SQUARES SQUARE F VALUE PROB>F
MODEL 2 0017931093 0018965464 0.571 0.5695
ERROR 40 0.62814987 0.01570375
C TOTAL 42 0.64608080
ROOT MSE 0.1253146 R-SQUARE 0.0278
DEP MEAN 0.3534683 ADJ R-SQ -0.0209
C.V. 35.45285
PARAMETER ESTTMATES
PARAMETER STANDARD T FOR HO:
VARIABLE DF ESTIMATE ERROR PARAMETER=0 PROB > ITI
INTERCEP 1 0.33828241 0.03240362 10.440 0.0001
QUART 1 0.01199379 0.02052322 0.584 0.5622
BECK 1 0.003103492 0101962563 0.625 0.5353
Table 33A. FEMALE ONLY T-TEST
DEP VARIABLE: CRIT_PCI‘
ANALYSIS OF VARIANCE
SUM OF MEAN
SOURCE DF SQUARES SQUARE F VALUE PROB>F
MODEL 2 0.009227354 0034613677 2.454 0.0988
ERROR 40 0.07521138 0.1!)1880284
C TOTAL 42 0.08443873
ROOT MSE 0.04336225 R-SQUARE 0.1093
DEP MEAN 0.03562748 ADJ R-SQ 01547
C.V. 121.7101
PARAMETER ESTTMATES
PARAMETER STANDARD T FOR H0:
VARIABLE DF ESTIMATE ERROR PARAMETER=0 PROB > ITI
INTERCEP 1 0.01833245 0.01121253 1.635 0.1099
QUART 1 0.001776648 0.1!)7101590 0.250 0.8037
BECK 1 0.003373%? 0.%1717181 1.965 0.0564

 

73

Table 34A. FEMALE ONLY T-TEST
DEP VARIABLE: QUES_PCT
ANALYSIS OF VARIANCE
SUM OF MEAN
SOURCE DF SQUARES SQUARE F VALUE PROB>F
MODEL 2 0.05508122 0.02754061 1.340 0.2735
ERROR 40 0.82236916 0.02055923
C TOTAL 42 0.87745037
ROOT MSE 0.1433849 R-SQUARE 0.1528
DEP MEAN 0.5718319 ADJ R-SQ 0.0159
C.V. 25.07466
PARAMETER ESTIMATES
PARAMETER STANDARD T FOR H0:
VARIABLE DF ESTIMATE ERROR PARAMETER=0 PROB > ['11
INTERCEP 1 0.59545672 0.03707621 16.150 0.0(X11
QUART 1 0.02325434 0.02348266 -0.990 0.3280
BECK 1 0004890220 01115678162 -0.861 0.3942

 

74

Table 35A. HAMILTON DEPRESSION SCORE DISTRIBUTION

HISTOGRAM #
4754-" 1
32.54! 1
:1“: 6
17.5+"'"""" 7
.8100“... 16
.tttttttttttttttttttttt 43

.A AAAAAA—LAJ .4 AAAAAAAA A4..- 444444444444.- .. .A 4.- .- .A.‘ 4.;
T‘v—TT-v—T'r-v-v-Tvvvw—TvV—TTTTTTTw-TT—r—rv-v-vvvvTvaTva-V-

 

87

_1_ EL .1. .1. _L .L .1. _L
U

" MAY REPRESENT UP TO 2 COUNTS

 

Table 36A. NORMAL PROBABILITY PLOT FOR HAMILTON DEPRESSION SCORES

475+ *

I

I
32.5-1- "

l I

' .‘OOOW
175+ *“++++

I Httttttt

l +tttttlttl

zatttlttlfiltIlltlltiltlttfifi
+ L + 4 + + + + + + +

U I

-2 -1 0 +1 +2

75

Table 37A. BECK DEPRESSION SCORES DISTRIBUTION

STEMLEAF #
190
18
17
160 1
15

14

130

12000

11001!)

10000

90000

8000w0000

7000000

6000m000

500000000011100000

4W

300000000000000000000
ZOOOWOOOOMOOOOO
IOOOMOOOOOOOOOOWOOOO
OOOOIDOOOOCXXWWOOOOWOOOONO

T I V U V T

0—0

 

N
**8—5858m6~65w3u_

_L
U

«r

76

Table 38A. NORMAL PROBABILITY PLOT BECK DEPRESSION INVENTORY SCORE

 

19.5+ *
I
I
I t
I
|
I I
I it! H
l ** «14+
I ** H-
I "HI-
I titH
I "+
I “'4-
' 00!.
I +1..
I tilt
I til.
I ##tt
0.5+tttlttltfittttltlit
+ + + + + + + + + + +

-2 -l 0 +1 +2

77

hlchigan §tate University Familz Study
East Lansing. Michigan 48824-1117

RESEARCH PARTICIPATION INFORHED CONSENT FORM

we freely consent to take part In a scientific study of child
development and Family functioning being conducted by Ors. Robert A.
Zucker, Robert B. Noll and Hiram E. Fitzgerald (From the Departments of
Psychology 8 Pediatrics and Human Development).

The study has been explained to us and we understand the explanation
that has been given and what our participation and what the participation

 

of our chiid(ren) will involve. He understand
that we will be asked to Fill out questionnaires. and take part In
interviews. and that our childiren) will do the same. He have been
informed that participation in the project involves taking part in 9
sessions with a time commitment of approximately II hours for each adult

and approximately 7 hours for our child.

We understand that we are Free to discontinue our participation in
the study at any time without penalty: we also are aware that our
involvement or lack of involvement In the study. or in different subparts
ofir. will not affect our obtaining treatment that Is offered by your
staff in any other capacity.

we understand that the results of the study will be treated in
strict confidence and that we and our childiren) will
remain anonymous. within these restrictions. results of the study will
be made avaiiaOIe to us periodically throughoOt the cOurse OF the project
and For a minimum or 3 years after the project has concluded. Also
w‘thin the restrictions noted above. we understand that general results
of the research will appear in professional journals and will be
presented at scientific meetings.

we understand that our participation in the study does not guarantee
any beneficial results to us or to members of our family.

 

 

 

 

Signed:
Hother Father
Child (when appropriate) Date

 

 

witness Family sutiJect number

78

We underetand that our participation in the atudy dpee not
:ua::ntee any beneficial reaulta to ua or to membere of our
amy.

 

 

 

 

 

 

 

Signed:
E
Mother Father 7
F
Child (when appropriate) Date K
Witneaa Family eubject number

9/89

79

DATA RECORDING SHEET

Respondent Number:

 

Test: 71.0 72.0

hother Father CDI

 

 

Pgrgn; fighgvigrs

T3.0

 

 

Acknowledge

Tbtlfi

 

Irrelevant
verbalizatlon

 

Critical
Statement

 

Physical
Negative

 

Physical
Positive

 

Unlabeled
Praise

 

Labeled
Praise

 

Desc/Refl
Question

 

Reflective
Statement

 

Descriptive
Statement

 

Indirect Command followed by:

No Opportunity

Compliance

Noncompliance

ICC...

 

 

 

 

 

Given By:
Date:
Clean-up Time
il h viors Yoga;
1
Changes
Activity
Cry
Ignored

ROSDONOCG t0

 

Yell
Ignored

Responded to

 

Rhine
Ignored

RCSDOflde t0

 

Smart Talk
Ignored

Responded to

 

Destructive
Ignored

...... OOOOOOOOOOQOOOOOOOOO...

RISDOfldOd t0

 

Physical Negative
Ignored

Responded to

 

 

Direct Cmnd followed by:

No Opportunity

Compliance

Noncompliance

 

Other

 

 

Other
Ignored

 

ReSOOMed to

 

 

 

 

REFERENCES

Aragona, J. A., and Eyberg, S., (1981). Neglected Children: Mother’s Report
of Child Behavior Problems and observed verbal behavior. th_d
Development, 52, 596-602.

Beardslee, W., Bemporad, 1., Keller, M., and Klerman, G. (1983) Children Of
Parents with Major Affective Disorder: A Review. Thg American Jomal of
Psychiag. 140: 825-832.

Beck, A.T. (1971). Cognition, Affect, and Psychopathology. Agghives if
mega; Psychiag, 24: 295-500.

Beck, A.T., & Beck, R.W. (1972). Screening depressed patients in family
practice: A rapid technique. Postmuate Mediging, 52, 81-85.

Beck, A.T., Steer, R.A., & Garbin, MG. (1988). Psychometric properties of the
Beck Depression Inventory: Twenty-five years of evaluation. QM
Psychology Review, 8, 77-100.

Bright, M., & Stockdale, D. (1984) Mothers’, fathers’, and preschool Children’s
interactive behaviors in a play setting. M if Mi; Mg, 144,

219-232.

80

81
Brown, 6., Bhrolchain, M., & Harris, T. (1975). Social class and psychiatric

disturbance among women in an urban population. May, 9 (2), 225-
254.

Clarke-Stewart, A. (1973). Interactions between mothers and their young
children: characteristics and consequences. Monomhs of the Smigg for
Rgggh in Child Development, 38(6-7, Serial No. 153).

Clarke-Stewart, A. (1978). And daddy makes three: the father’s impact on
mother and young child. M W 49, 466-478.

Clarke-Stewart, A. (1980). The father’s contribution to children’s cognitive
and social development in early childhood. In Frank A. Pedersen, (Ed), Ih_e

Pager-Infant Relatignship. New York: Praeger.
Colletta, N. D. (1983). At risk for depression: A study of young mothers. 1h;

19%;; 9f fined; Psychology, 142, 301-310.

Cox, A. D., Puckering, C., Pound, A., & Mills, M. (1987). The impact of
maternal depression in young children. M if gm Psychglggy Qfld
Psychiat_ry, 28(6), 917-928.

Davies, W. H., Zucker, R. A., Noll, R. B., & Fitzgerald, H. E. (1991). Early
socialization practices in alcoholic families: The relationship of child-rearing
patterns to demographics and parental psychopathology. Michigan State
University, E. Lansing, MI.

Downey, G., & Coyne, J.C. (1990). Children of depressed parents: An

integrative review. Psyghglog'ggl Bulletin, 108(1), 50-76.

 

82
Dumas, J. E., Gibson, J. A., & Albin, J. B. (1989) Behavioral correlates of

maternal depressive symptomatology in conduct-disorder children. MIMI 9_f
ansulting g4 mam-114193 57(4), 516-521.

Easterbrooks, M. A. & Goldberg, W. A. (1984). Toddler development in the
family: Impact of father involvement and parent characteristics. Child
Wt. 55, 740-752.

Eyberg, S. (1978). Assessment of child behavior problems: The validation of
a new inventory. Mfimmlfim S 78, 113-116.

Eyberg, S., and Matarazzo, R. G. (1980). Training parents as therapists: A
comparison between individual parent-child interaction and parent group
didactic training. M 9_f_‘ M gsthnggy, 36, 492-499.

Eyberg, S., and Robinson, E. A., (1983). Dyadic Parent Child Interaction
Coding System: A Manual. W W 13, 24. (Ms. No.
2582).

Fawl, C. L. (1963). Disturbances experienced by children in their natural
habitat. In R. Barker (Ed), m M 21’ Behavior. New York: Appleton-
Century-Crofts.

Fergusson, D. M., Horwood, L. 1., Gretton, M. E., & Shannon,

F. T. (1985). Family life events, maternal depression,
and teacher descriptions of child behavior. Magic;

75(1), 30-35.

83
Fitzgerald, H. E., Sullivan, L. A., Ham, H. P., Zucker, R. A., Bruckel. S.,

Schneider, A. M., & Noll, R. B. (under review). Predictors Of behavioral
problems in three-year-old sons Of alcoholics: Early evidence for the onset
of risk. Michigan State University, East Lansing, MI.

Forehand. R., & Brody, G. (1985) The association between parental personal]
marital adjustment and parent-child interactions in a clinic sample.
thaviogg Resea_rgh 1134 My, 23, 211-212.

Forehand, R., Lautenschlager, G. J., Faust, J., & Graziano, W. G. (1986)
Parent perceptions and parent-child interactions in clinic-referred children: A
preliminary investigation of the effects of maternal depressive moods.
mmmm 24. 73-75-

Goldman, N., & Ravid, R. (1980). Community surveys: Sex differences in
mental illness. In M. Guttentag, S. Salasin, and D. Belle (Eds). mm
m if m. New York: Academic Press.

Grunebaum, H., Cohler, B., Kauffman, C., & Gallant, D. (1978). Children of
Depressed and Schizophrenic Mothers. Q1114 Psyc_hia_tgy .a_nd_ Human
Development, Vol.8 (4), 219-228.

Hamilton, M. (1960). A rating scale for depression. 19mg; 61m
W and Emblem. 23. 56-62-

Hamilton, M. (1969). Standardized assessment and recording of depressive

symptoms. mmmm 72. 201-205.

 

84
Hayghe, H.V. (1990). Family members in the work force. Monthly Lam

m 113, 14-19.

Hopkins, 1., Marcus, M., & Campbell, S. (1984). Postpartum depression: a
critical review. Psychologicg Mega, 95(3), 498-515.

Jacob, T. & Leonard, K. (1986). Psychosocial functioning in children of alcoholic
fathers, depressed fathers, and control fathers. M if m m ;A_l@h_o_l, 47(5),
373-380.

Kotelchuck, M. (1976) The infant’s relationship to the father: experimental evidence.
In M.E. Lamb (Ed), m 13:11; 91’ the _Fa_thg E Qh_il_d Development. New York:
Wiley.

Lamb, M. E. (1976). The role of the father: An overview. In M.E. Lamb (Ed), fl;
mammamm New York: Wiley-

Longfellow, C., Zelkowitz, P., & Saunders, E. (1982) The Quality of Mother Child
Relationships. In Lam Q m D. Belle, (Ed), Beverly Hills: Sage Publications.

McKnew, D. H., & Cytryn, L. (1973) Historical background in children with affective
disorders. American M o_f Psychiag, 130, 178-180.

Mueller, C. W. & Parcel, T. L. (1981). Measures of
socioeconomic status: Alternatives and recommendations.
cm W 52, 13-30.

Nolen-Hoeksema, S. (1987). Sex differences in unipolar depression: Evidence and

theory. mm agnggn, 101 (2), 259-282.

 

85
Nyquist, L., Slivken, K, Spence, 1., & Helrnreich, R. (1985). Household

responsibilities in middle class couples: The contribution of demographic and
personality variables. _S_e; gm 12(1/2), 15-34.

Orvaschel, H., Weissman, M., & Kidd, K. (1980) Children and depression: the
children of depressed parents; the childhood of depressed patients; depression in
children. m of Affective mm, 2, 1-16.

Patterson, G. (1976). The aggressive child: Victim and architect of a coercive system.
In E. J. Mash, L. A. Harnerlynck, & L. 0. Handy (Eds.), Behavior mgifigtign g1,
My; Vol. 1. 113m mm. New York: Brunner/Mazell, 1976.

Patterson, G. (1981). Mothers: the unacknowledged victims. mm of the
mmMamm 136. (45). 5-

Pearlin, L., & Johnson, J. (1977). Marital status, life strains, and depression.
WWW 42, 704-715.

Radloff, L. (1975). Sex differences in depression: The effects of occupation and
marital status. mm A M 91' my, 1, 249-266.

Regier, D. A., Farmer, M. E., Rae, D. S., Locke, B. 2., Keith, S. J., Judd, L. L., &
Goodwin, F. K. (1990). Comorbidity of mental disorders with alcohol and other
drug abuse. Mgfmmmm 264 (19), 2511-2518.

Robinson, E., and Eyberg, S. (1981). The Dyadic Parent-Child Interaction Coding
System: Standardization and Validation. M if (insulting and (meal

mm 52. 245-250.

86
Robinson, E. A., and Eyberg, S. M. (1984). Behavioral Assessment in pediatric

settings: Theory, method and application. In P. R. Magrab (Ed). W Ed.
mm Asgessmeng: M m M 9m. New York: Plenum.

Rogers, T. R., & Forehand, R. (1983) The role of parental depression in interactions
between mothers and their clinic-referred children. Comm Therapy and Rese_agh,
17, 315-324.

Rutter, M. (1966) calm 1f Sick Megs. An Enm’ nmentg and Psychiag'g M.
Institute of Psychiatry, Maudsley Monograph 16. London: Oxford University Press.

Stevens, G., & Feathennan, D. L. (1981). A revised socioeconomic index of
occupational status. m M 3m 10, 364-395.

Susman, E. J., Trickett, P. K., Iannotti, R. J., Hollenbeck, B. E., & Zahn- Waxler, C.
(1985). Childrearing Patterns in Depressed, Abusive, and Normal Mothers.
Amman Jamal S! W 55 (2). 237-251.

Webster-Stratton, C., & Hammond, M. (1988) Maternal depression and its
relationship to life stress, perceptions of child behaviors, and child conduct
problems. Lourn__al_ of Abnormal CM Psychology, 16(3), 299-315.

Webster-Stratton, C. (1988). Mothers’ and fathers’ perceptions of
child deviance: Roles of parent and child behaviors and parent
adjustment 19m 2f mm m Slim amen.

56(6), 909-115.

 

87
Weissman, M. M. (1980). The treatment of depressed women: The efficacy of

psychotherapy. In C. L. Heckerman (Ed), _'l_h_o mm , (p.307-324.

New York: Human Sciences Press.

Weissman, M. M. & Klerman, G. (1977). Sex differences and the epidemiology of
depression. Archives of General Psychiatry, 34, 98.

Weissman, M. M., and Paykel, E. S. (1974). mwm. Chicago:
University of Chicago Press.

Weissman, M. M., Paykel, E. S., & Klerman, G. (1972). The Depressed Woman as a
Mother. M Mom, 7, 98-108.

West, M. 0., & Prinz, R. J. (1987). Parental alcoholism and childhood
psychopathology. Psychology“ al Mm 102, 204-218.

Williams, H., & Carmichael, A. (1985). Depression in mothers in a multi-ethnic urban
industrial municipality in Melbourne. Aetiological factors and effects on infants
and preschool children. logo; of thd Won mo Wm 26, (2), 277-
288.

Yogman, M. W., (1982). Development of the father-infant relationship. In H.
Fitzgerald. B. Lester, & M. W. Yogman, (Eds.). 111m mm for
Behavigfl Poor'atr'ios, Vol. 1. New York: Plenum.

Zelkowitz, P. (1982). Parenting philosophies and practices. In D. Bell, (Ed) Li_ve_s_ in
m: Moo m m (pp. 154-162). Beverly Hills: Sage Publications.

Zucker, R.A., Noll, R., & Fitzgerald, HE. (1986). Grant proposal: Risk and coping in

children of alcoholics. Michigan State University, East Lansing, MI.

88
Zucker, R. A. (1991). Scaling the developmental momentum of alcoholic process via

the Lifetime Alcohol Problems Score. Alcohol god Alcoholism, suppl. 1, 505-510.

Pergamon Press: Great Britain.

"7'1? llllllllllllllll“