 

 

r £1673) #1;
51.13.!!! I.

. 7.1.1
at: iii-1:]... 3

.r ..
it): 1:
’25:); {ill}?! (310...?
» .)I.ioi)ll...

.13) I331}. $1111.}

 

 

1 7!.)51’;
)i 10......11

(5:1.

‘5}, (in! Alu.5)li'(olrtllut1;. .. u

53 1"...)‘la‘fa
I: .1. a 4 112.415... link

 

 

 

 

 

£fr1§$i5

SIIRATYLB

i||||ﬂ|l|||||tlll“IHHHIWNWWIN“ ||||L|||||U|

3 1293 00885 329

 

 

 

 

 

.3 i3 drink to certify that the

-- -.._...-_-—-a.—-—
~m--~-- p...

dissertation'entitled

Parent and Early Adolescent Communication
in Two-Parent Families:
An Ecoiogical Study

presented by

Christine Joyce Thompson

has been accepted towards fulﬁllment ‘
of the requirements for

PhD Family Ecoiogy

degree in

 

 

1L4L4«£zr/éy7%L£hL2;:;

Major profess!”

Date 8/3/673

MSU is an Afﬁrmative Anion/Equal Opportunity Instilulirm 0-12771

 

 

ETUQN BOX to remove this checkout from your record.

P M35010 FINES return on or before date due.
r0 A

WOUE DATE DUE DATE DUE

 

Yﬂﬁmtﬂ
111409

 

 

 

 

 

 

 

 

 

 

 

JAN717 2

 

MAR 2 1 2005s

i;
3,

 

 

 

 

 

 

 

MSU Is An Affirmative Action/Equal Opportunity Institution

c:\c|rc\dateduerpm3-n.[

 

PARENT AND EARLY ADOLESCENT COMMUNICATION
IN TWO-PARENT FAMILIES:
AN ECOLOGICAL STUDY

BY

CHRISTINE JOYCE THOMPSON

A DISSERTATION

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

DOCTOR OF PHILOSOPHY

Department of Family and Child Ecology

ABSTRACT

PARENT AND EARLY ADOLESCENT COMMUNICATION
IN TWO-PARENT FAMILIES:
AN ECOLOGICAL STUDY

By

Christine Joyce Thompson

This study examines communication between parents and their early
adolescent children from an ecological perspective to determine the relative inﬂuence
of speciﬁc variables on communication between parents and early adolescents in two-
parent families. The impact of selected physiological, sociological and psychological
variables on intrafamily communication ratings was examined. The independent
variables examined were: gender and pubertal age of child, work status of parents,
educational level of mother and father, family income, parents’ perceptions of family
stress and job-parent role stress, and the early adolescents’ perceptions of acceptance
by mother and father. The outcome variables were parents’ and early adolescents’
ratings of each other using the Parent-Adolescent Communication Scale (Olson,
1982), including problems and openness subscales.

The sample used consisted of 74 two-parent families--mother, father and early
adolescent (10 to 14 years old), who participated in a state-wide survey in 1987.
Youths were interviewed in their homes while parents completed a questionnaire.

Subjects were selected using a stratiﬁed multi-stage cluster sampling technique.

[l

 

Ordinary Least Squares multiple regression was used to examine the relative
importance and direction of inﬂuence of the independent variables on
communication ratings. Of the eight regression equations examining mothers’ and
fathers’ ratings of early adolescents and the youths’ ratings of their parents on both
the openness and problem-free subscales, six were statistically signiﬁcant. In general,
physiological variables were important factors in early adolescents’ ratings of parents
on openness. Pubertal age was negatively related to the early adolescent ratings of
both mother and father on openness. Psychological variables were important factors
in both early adolescent and parental ratings. The early adolescents’ perception of
acceptance by mother and father was found to be positively and signiﬁcantly related
to communication ratings in six of the eight equations. Fathers’ parent-job role stress
had a negative impact on parent-child communication, whereas mothers’ parent-job
role stress was not found to be signiﬁcant in this sample of two-parent families.
Family stress was found to be negatively related to communication ratings; stress was
found to be a greater factor in mothers’ ratings than in fathers’ ratings of children.
Fathers’ education was positively related to communication ratings in the father-child
dyad; mothers’ education was not signiﬁcant to ratings in the mother-chﬂd dyad.
Family income was not statistically significant in explaining differences in

communication ratings.

ACKNOWLEDGEMENTS

There are many persons who have encouraged, guided and assisted me in my
Ph.D. journey. I would like to express thanks and appreciation to a few of those
special people:

To Mary Andrews, not only for serving on my doctoral committee, but also
for introducing me to the process of research and helping me grow to recognize my
own capabilities;

To Rosemary Walker and Dennis Keefe, committee members whose
assistance, support and thoughtful evaluation are much appreciated;

To Joanne Keith, chairperson of my doctoral committee--her guidance,
support and patience were extremely helpful to me in completing my degree after
moving away from Michigan State University;

To Judy Pfaff, David Sommers and Xiongwen Rui for assistance with data
analysis and statistical advice;

To my husband, Stan, for all of his support, advice, statistical consultations
and encouragement-~the completion of this dissertation would not have been possible

without his help.

iv

TABLE OF CONTENTS

LIST OF TABLES ............................................. vii
LIST OF FIGURES ........................................... viii
CHAPTER 1 - INTRODUCTION .................................. 1
Background of Problem ......................................... 1
Purpose of the Research ........................................ 4
Conceptual Framework ......................................... 6
Theoretical Framework ......................................... 6
CHAPTER 2 - LITERATURE REVIEW ........................... 12
Gender of Early Adolescent ................................... 13
Pubertal Age of Early Adolescent ............................... 14
Parental Employment Status ................................... 16
Income and Education ........................................ 18
Parent-J ob Role Stress ....................................... 22
Perception of Family Stress by Mother and Father ................... 25
Child’s Perceived Acceptance by Mother and Father ................. 27
Within Family Differences ..................................... 28
CHAPTER 3 - METHODOLOGY ................................ 30
The Data/Sample ............................................ 30
Sample Selection ............................................. 31
Sample Description ........................................... 33
Measurement Procedures ....................................... 33
Instruments ................................................. 36
The Statistical Model .......................................... 43

Hypothesized Relationships ..................................... 47

Physiological ............................................... 48
Psychological ............................................... 48
Sociological ................................................ 49
Overview of Statistical Methods .................................. 50
Limitations of the Study ........................................ 52
CHAPTER 4 - ANALYSIS AND FINDINGS ........................ 55
Overview of Analysis .......................................... 55
Examining the Data for Iinearity/Multicollinearity .................... 56
Test Results for Cross-Equation Correlations and
Simultaneity ............................................... 57
Descriptives ................................................. 67
The Dependent Variables ..................................... 67
Independent Variables ........................................ 70
Multiple Regression Results ................................... 72
Summary of Findings .......................................... 81
Within Family Correlations ..................................... 85
CHAPTER 5 - CONCLUSIONS AND IMPLICATIONS ................ 86
Conclusions ................................................. 86
Physiological Factors ......................................... 87
The Sociological Factors ...................................... 89
The Psychological Factors ..................................... 91
Within Family Differences ..................................... 93
Implications ................................................. 94
Implications for Research ..................................... 94
Implications for Education ..................................... 96
Implications for Family Policy .................................. 98
REFERENCES ............................................... 99
APPENDIX ................................................. 108

LIST OF TABLES

Table 3-1. Total Sample Compared to Multiple Regression
Subsample: Selected Demographics ............................... 34

Table 3-2. Overview of Item and Scale Formulations
by Variable ................................................. 38

Table 3-3. Overview of Research .................................. 44

Table 34. Eight Equation Statistical Model of Parent-
Child Communication ......................................... 45

Table 4-1. Pearson Correlation Coefﬁcients Parent-
Child Communication Variables .................................. 5 8

Table 4-2. Items Measuring Parent-Chﬂd Communication
Openness: Percent of Mothers, Fathers and Children
Agreeing with Each Statement ................................... 60

Table 4-3. Items Measuring Parent-Child Communication
Problems: Percent of Mothers, Fathers and Children
Agreeing with Each Statement ................................... 61

Table 4-4. Parent-Child Communication Statistical Model:
Ordinary Least Squares Results .................................. 63

Table A-1. Parent-Chﬂd Communication Statistical Model:
Seemingly Unrelated Regression (SUR) Results ..................... 108

vii

LIST OF FIGURES

Figure 1-1. Variables and Levels of Analysis Examined in
Parent-Chﬂd Communication Study ................................ 9

Figure 4-1. The Relationship Between Early Adolescents’
Perceived Acceptance by Mother and Father and
Communication Openness ...................................... 74

viii

CHAPTER 1

INTRODUCTION

BACKGROUND OF PROBLEM

Communication, the symbolic, transactional process of creating and sharing
meanings, is central to family functioning and human development within the family.
Communication between parents and adolescents is important to the psychosocial
development of the adolescent. Parental behaviors of enabling (communicative
behaviors exhibiting warmth, support and problem solving) have been found to be
positively related to adolescent ego development whereas constraining behaviors
(judging, devaluing, blaming) have been found to be negatively related to ego
development (Hauser, Powers, Noam and Jacobson, 1984).

Communication has been shown to be a facilitator of other important family
systems characteristics. Communication, adaptability and cohesion are major

components of family interaction. In their Circumplex Model of Families, Olson,

3'1

 

 

2

Russell and Sprenkle (1983), show that communication is a facilitator of cohesion
and adaptability within the family system.

Communication within families is a dynamic process, changing over time as
the children grow from dependent newborns to autonomous adults. Families differ
with respect to the nature of their interpersonal relationships, depending upon
individual needs, desires and expectations as well as members’ interpretation of
cultural values and norms. Communication is essential to the negotiation processes
families use to meet the developmental changes dictated by the growth of individual
members.

Early adolescence has been identiﬁed as problematic with respect to parent-
child relations. Parents report that ages 12 and 13 are the most difﬁcult in terms of
parenting (Offer, 1969). Longitudinal studies have found that there is improvement
in parent-adolescent relations in grades 10 and 11, suggesting that early adolescence
is a low point in the relationship (Hill and Holmbeck, 1987; Hill, Holmbeck, Marlow
and Lynch, 1985; Steinberg, 1987; Collins, 1990). Montemayor (1983) found that
conﬂict is normal between parents and early adolescents, even in reportedly "close"
relationships. As children grow toward adulthood, parallel changes are needed in
their relationship with their parents to facilitate this transition, or at least to remove
obstacles to the demands of the developmental tasks faced by early adolescents.

Difﬁculties in the communicative relationship between parents and early
adolescents is due, in part, to the adolescents’ need for autonomy. Early adolescents

seem to have a need to engage in conﬂict with parents over seemingly unimportant

issues (Hill, 1980; Montemayor, 1983). There is evidence, however, that parents also
play an important role. Early adolescence is a transitional time, with parents
gradually moving from controlling their children to granting their emerging
adolescents more autonomy (Offer, 1969; Hill, 1980). Some parents may exert too
much control for too long, adversely affecting the parent-adolescent relationship, and,
in some cases pushing the adolescent into antisocial behavior (Hill, 1980; Baumrind,
1968). In view of the importance of the parent-chﬂd relationship to the psychosocial
development of the adolescent and the changing nature of the parent-child
relationship beginning in early adolescence, it is important to understand the
communication patterns between early adolescents and their parents. An ecological
approach to the study of family communication, can help families identify
environmental factors that impact communication.

Despite the recognized importance of communication to family relationships,
research into the nature of family communication presents some challenging
difﬁculties. One of the main difficulties is the complexity of family communication
resulting in a variety of aspects upon which researchers might focus. Communication
has been studied in the following ways: as a general construct; as self-disclosure; as
a skill-learning process during therapy; in terms of different styles or patterns of
interaction; in terms of speciﬁc components such as empathy, congruence, and
regard; and in terms of couples’ hesitancy to communicate within the context of a

close interpersonal relationship (Olson, Portner and Bell, 1982).

4

Most studies in the area of communication have focused on communication
in the spousal relationship as opposed to intra-family communication (Olson, Portner
and Bell, 1982). The literature in the area of communication between parents and
children most often focuses on adolescence as opposed to early adolescence.
Differences between mothers and fathers in terms of their communicative roles with
their children have been examined, as have differences in patterns of communication
and closeness between mother and child versus father and child. However, there is
a dearth of research which attempts to identify the relative contribution of
environmental factors to the communicative relationship between parents and early
adolescents.

Clearly, the communicative relationship is not unidirectional, that is, with
adolescents creating problems for their parents, but rather, a bidirectional, interactive
relationship and should be studied as such. It should also be recognized that the
relationship is complex; physiological changes in early adolescence bring about
changes in the parent-chﬂd relationship, but consideration of social-psychological
issues are important for total understanding (Montemayor, Adams and Gullotta,

1990). A systems approach is most effective in studying this relationship.

PURPOSE OF THE RESEARCH

The purpose of this study is to examine communication between parents and

their early adolescent children from an ecological perspective to determine the

5
relative inﬂuence of speciﬁc physiological, sociological and psychological variables.

The questions to be addressed are as follows:

What is the role of early adolescent characteristics of pubertal age and gender
in the parent-chﬂd communicative relationship?

What is the role of family employment typologies (i.e., dual earneruboth full
time, dual earner-~one full time, one part time, and traditional-single earner and
homemaker) in the parent-child communicative relationship?

What is the role of stress associated with balancing work and family demands--
for both parents and early adolescents?

What is the role of family income and educational levels of both mother and
father in the parent-child communicative relationship?

What is the extent of inﬂuence of parental perceptions of family stress on the
parent-child communicative relationship?

What is the extent of inﬂuence of the early adolescents’ perceived level of
acceptance by their mothers and fathers in the parent-child communication
relationship?

Are the relative inﬂuence of the speciﬁed variables different for

communication between adolescents and fathers versus adolescents and mothers?

6
CONCEPTUAL FRAMEWORK

The human ecological approach to the study of the family provides the
conceptual framework for this study. Human ecology is concerned with the
interaction and interdependence of humans (as individuals, groups and societies) with
the environment (Bubolz and Sontag, 1990). An ecological approach to the study of
any living thing has three major components: the system (or organism), the
environment, and the interactions between the two (Bronfenbrenner, 1979). The
ecosystems perspective views families, as well as individual human beings, as
organisms interacting with their total environment. This environment is multilevel
and includes human as well as nonhuman elements. This study considers biological,

sociological and psychological factors.
THEORETICAL FRAMEWORK

The theoretical framework for this study is that of the developmental
icontextual view of human development (Lerner, Lerner and Tubman, 1990). This
theory is also referred to as "probabilistic epigenetic" and refers to the interacting of
biological, psychological, and sociological factors in human development and
behavior. ‘1‘

Traditionally, developmental psychology has focused on a biological model of

change. A strong link between personality conﬁgurations and physiological

7
functioning has been stressed by organismic theorists (Erickson, 1959 & 1968;
Sigmund Freud, 1949; and Piaget, 1970). In this type of theory, which as also been
labeled "predetermined epigenetic," biology is seen as the prime mover of
development. These organismic theorists’ views ﬁt with what has been termed a
"direct effects" model of the inﬂuence of organismic variables on adolescent and early
adolescent psychosocial functioning (Lerner et al., 1990). This model speciﬁes that
the changes associated with early adolescent biological change inﬂuence directly
psychosocial characteristics. The inﬂuences are unidirectional; that is, inevitable
biological changes bring about developmental disturbances that have psychosocial
repercussions.

In contrast, probabilistic epigenetic theory, which has also been labeled
"dynamic interactional" or "developmental contextual", assumes that organismic
changes both affect and are affected by the context in which they occur (Lerner et
al., 1990). The term "probabilistic epigenetic" was used by Gottlieb (1970):

to designate the view that the behavioral development of individuals

within a species does not follow an invariant or inevitable course, and,

more speciﬁcally, that the sequence or outcome of individual

behavioral development is probable (with respect to norms) rather

than certain. (p. 123).

A basic tenet of probabilistic epigenetic theory is that characteristics of organismic
individuality inﬂuence and are inﬂuenced by the psychosocial context within which
they are expressed. Thus, the examination of directional relationships among

biological, psychological and social processes is important in understanding

development and behavior.

“I’m—I‘mt'r ...

8

Bronfenbrenner (1979) can be viewed as a developmental contextual theorist in
that his human ecological theory speciﬁes that the developing organism (individual)
is inﬂuenced by the environments that surround him or her; the biological,
sociological, psychological, physical and economic characteristics of the environment
all impact the developing individual.

Family ecological theory (Bubolz and Sontag, 1990) is also a developmental
contextual theory; however, the family as opposed to the individual is speciﬁed as the
"organism" which interacts with a dynamic environment. Family ecological theory
assumes that the properties of families and the environment, the structure of
environmental settings, and the processes taking place within and between them must
be viewed as interdependent and analyzed as a system.

The Developmental Contextual Model of Person-Context Interaction (Lerner
et al., 1990), depicts a dynamically interactive, multilevel, model of human behavior
and development. The authors point out that it would not be possible or even useful
to engage in research that would test the model as a whole; however, it does provide
a framework for looking at the contextual inﬂuences on the adolescent as well as the
"adolescent effects" on the social context. The developmental contextual model was
used as a guide for this study in the selection of variables from the individual and
contextual levels depicted. In Figure 1-1 the model used for this study of the relative
inﬂuence of contributing factors to communication between parents and early

adolescents is shown.

9

COMMUNITY

   

FAMILY

INCOME

    

MOTHER

 

Figure 1-1. Variables and Levels of Analysis Examined in
Parent-Child Communication Study.

Dependent Variables

CMOMO = Early adolescents’ ratings of mother
(openness)

CMOMP = Early adolescents’ ratings of mother
(problem-free)

MCHTLKO = Mothers’ ratings of child (open-
ness)

MCHTLKP = Mothers’ ratings of child
(problem-free)

CDADO = Early adolescents’ ratings of father
(openness)

CDADP = Early adolescents’ ratings of mother
(problem-free)

DCHTLKO = Fathers’ ratings of child (open-
ness)

DCHTLKP = Fathers’ ratings of child (problem-
free)

Independent Variables

INCOME = Family income

PUBRTAGE = Pubertal age of child

GENDER = Gender of child

MACCEPT = Child’s perceived acceptance by
mother

DACCEPT = Child’s perceived acceptance by father
MOMEDUC = Mother’s education

DADEDUC = Father’s education

MFAMSTRS = Mother’s perceived family stress
WORKSTATUS = Parental work status, FF, FN, FP
DFAMSTRS = Father’s perceived family stress
MWKSTRES = Mother’s job-parent role stress
DWKSTRES = Father’s joboparent role stress

10

The current study focuses on family communication as reported by each of
three different family members--mother, father, and their early adolescent child.
When applied to the present study, the model calls for an examination of the relative
inﬂuence of speciﬁc physiological, sociological and psychological variables on
communication between parents and their early adolescent children. Communication
is measured by reports of mothers, fathers and early adolescents on communication
openness or self-disclosure between parent and child, and the extent of problems or
barriers to family communication, including the degree to which family members are
selective in their discussion with other family members.

The model speciﬁes consideration of the following inﬂuencing or independent
variables: gender and pubertal age of the child (physiological); work status and
education of parents as well as family income (sociological); parents’ perceptions of
family stress and job-parent role stress; and adolescents’ perception of acceptance by
mother and father (psychological). The outcome variables are those of parents’ and
early adolescents’ ratings of one another on openness and problem-free
communication (see Chapter 3 for operational deﬁnitions). A multilevel approach is
utilized including: individual variables (e.g., gender of early adolescent); family
variables (e.g. family income); and, variables outside of the immediate family system
which impact the family, e.g., the workplace. A bidirectional approach considers the
inﬂuence of the adolescent’s characteristics on parent-child communication as well
as the inﬂuence of speciﬁc parental variables, such as the stress involved in

combining work and parenting responsibilities. Psychological factors (i.e., the youths’

11

perception of acceptance by mother and father) are examined from the standpoint
of their contribution to adolescent ratings of parents as well as parental ratings of
adolescents.

To summarize, communication within families is generally recognized as
crucial to the development of individuals as well as to family dynamics of cohesion,
adaptability and problem-solving. For the most part, studies focusing on parent-chﬂd
relations during the early adolescent (as well as the adolescent) stage of development
are not ecological in their approach; they tend to focus on respective roles of
mothers and fathers with respect to speciﬁc communication constructs. In contrast,
the present study, based on developmental contextual theory, focuses on the role of
speciﬁc environmental factors to communication among family members. For
example, the relative impact of father’s and mother’s education on parent-chﬂd
communication is examined and compared for mother and father. In addition, self-
report ratings of family members are compared in order to determine differences in
perceptions of mother, father and early adolescent. For example, early adolescent
perceptions of communication or compared with parental perceptions in order to
gain a greater understanding of what can be expected during the transitional and
sometimes turbulent years of early adolescence. The ﬁndings of this study can be
used by researchers and family practitioners to help families better understand and

facilitate healthy communication with the family.

CHAPTER 2

LITERATURE REVIEW

In Chapter 1, the importance of the role of communication to the psychosocial
development of the individual as well as to family interaction processes (i.e.,
adaptability, cohesion and problem-solving) has been discussed. Communication is
generally accepted as one of the most crucial facets of interpersonal relationships.
Its prominence in the theoretical construction of human and, speciﬁcally, family
interactions attests to the great importance attributed to the role of communication.

Notwithstanding its importance, family communication is a difﬁcult and
challenging area for research. The multiple and varied aspects of family
communication have resulted in a myriad of communication concepts, constructs and
theoretical approaches within the literature; few studies are directly comparable.
However, a careful review of the literature yields some generalizations which will be
addressed in this chapter. The literature review is organized according to the
independent variables included in the study: pubertal age and gender of the early

adolescent; parental work status; family income and education of parents; parent-job

12

13

role stress; parental perceptions of family stress; and, child’s perception of acceptance
by mother and father. Findings vis-a-vis differences among family members with

respect to the way that they view communication in the family is also discussed.

GarderofEarlyAdolescent

Are there differences between male and female adolescents in communication
patterns and perceptions of communication with mothers and fathers? Barnes and
Olson (1985) found no differences between adolescent males and females in how
they perceived their communication with their mothers and fathers. Some
researchers have found that adolescent females disclose more and are disclosed to
more than males (Dimond and Munz, 1967; Kraft and Vraa, 1975; Yang and Hwang,
1980). N oller and Callan (1986) found that females of all ages reported talking more
often with mothers than did adolescent males. Adolescent males reported talking
more often than did females with fathers about interests, sexual issues, and general
problems. Davidson (1980) found that females’ disclosure scores were high in the
areas of general and personal information, whereas males’ scores were highest on
revealing sexuality to parents. Other investigators have found that male adolescents
disclose more to their fathers than do female adolescents (Balswick and Balkwell,
1977; Wiebe and Williams, 1972). Others studies indicate that although there are
gender differences in both the amount of disclosure and target of disclosure, speciﬁc
effects are unclear (J ourard, 1971; Komarovsky, 1974). Richardson, Abramowitz,

Asp and Petersen (1986), in examining 12 to 14-year-olds’ relationship with their

ﬂ 7 7,, ,Vm.uh Mwﬂ h: ._ —’_#—M :L’" ,

14

parents, found that girls in middle-class families reported better relationships with
their mothers than did boys. Grotevant & CDOper (1985) found that communication
variables related to identity exploration not only differed for males and females, but
also were signiﬁcant for them in different family relationships. For girls,
communication patterns in all four relationships were associated with identity
exploration, whereas for boys, only father-son interaction patterns were related to
exploration ratings. These differences suggest that sources of family inﬂuence on
identity exploration may be more diverse for female adolescents than for males.
Although differences according to gender have been found for speciﬁc
communication constructs, patterns and parental roles, it is unclear whether or not
the relative contribution of environmental factors affecting communication are

different for female and male early adolescents.

Pubertal Age of Eddy Adolescent

In examining age as an independent variable in parent-child communicative
relationships, no clear-cut patterns emerge from previous research. Rivenbark
(1971), in comparing children in the fourth, sixth, eighth, tenth, and twelfth grades,
found that boys’ disclosure to parents decreased signiﬁcantly around the tenth grade,
but returned to earlier levels by the twelfth grade. Girls’ disclosure to parents
remained about the same throughout adolescence.

In examining adolescent development, as opposed to age alone, an interesting

pattern emerges. Rivenbark (1971) has provided data on the developmental trend

15

of self-disclosure among adolescents to both parents and peers. An inverse,
curvilinear relationship was found between the physical development of adolescents
and their amount of self-disclosure to parents. The greatest amount of conﬂict in
the parent-adolescent relationship has been found to be at the peak of pubertal
growth--for both boys and girls (Steinberg and Hill, 1978; Steinberg, 1981; Norell,
1984). Disclosure to parents by adolescents decreased when the most obvious
physiological change-~the growth spurt--occurred. This decrease in disclosure was
most pronounced for boys. In studying seven-grade boys, researchers found
signiﬁcant relations between pubertal status and family interaction outcome variables
for the mother-son but not for the father-son dyad (Hill et al., 1985). These results
indicate that temporary perturbations occur in mother-son relations at or near the
peak of pubertal growth.

In examining affective family relations, Papini and Selby (1986), found that
differences among families according to the pubertal stage of the adolescent may be
the result of differences in father-adolescent and mother-adolescent relationships.
That is, affective differences between prepubertal and postpubertal families are
characterized by: (a) increased concern by the mother about her ability to control the
behavior of the postpubertal adolescent relative to the prepubertal adolescent; and

(b) increased dissatisfaction by the father with postpubertal adolescent behavior

relative to prepubertal adolescent behavior. It was found that relative satisfaction _

with family relationships decreases during transpuberty relative to prepuberty and

postpuberty, suggesting that the differences observed among families across all levels

16

of adolescent physical maturity are related to dialectic transformations of mother-
adolescent and father-adolescent relations. These results suggest that the peak of
pubertal growth, which may come at different ages for different children, is especially

problematic with respect to parent-child relations.

Parental Employment Status

Although the bulk of the research on parental work and parent-child relations
has been concerned with young children, inroads are being made into the area of
parental work and parent-child relations in adolescence and early adolescence. This
work has focused on maternal employment and on work status and numbers of hours
worked as opposed to psychosocial factors. Orthner (1990) suggests that it is
insufﬁcient to assess parental work solely in terms of whether the mother is employed
or not, or even in terms of the hours of employment, although the latter has usually
been ignored. It is also important to look at demands of the work environment, work
responsibilities, morale and job stress.

With respect to maternal employment status, an early study (Douvan &
Adelson, 1966) reported more positive parent-child relationships in families where
mothers were employed part time than in families where mothers were not employed
or were employed full time. Mothers employed part time reported greater closeness
to children than mothers in any other employment category. Hoffman (1989) has
shown that the relationship between amount of time mother spends on the job and

her child’s adjustment is not a linear one.

17

In examining the effects of maternal employment status on closeness between
parents and their early adolescent children, differences by gender of the adolescent
were reported (Paulson, Roman and Hill, 1990). Speciﬁcally, sons reported greater
closeness with fathers when mothers were employed and daughters reported greater
closeness with parents when mothers were not employed or were employed part time.
Mothers reported greater closeness to sons when they (mothers) were employed full
time or part time (part time was the highest); they reported greater closeness to
daughters when they were not employed or were employed part time. Fathers
reported no differences in closeness to either sons or daughters regardless of
maternal employment status. In a study of 64 tenth-grade adolescents, Montemayor
(1984) found that maternal employment status was related to adolescents’ relations
with both mothers and fathers. Males, but not females, were found to have more
arguments, which were of longer duration and greater intensity, with their mothers
when mothers worked than when they did not. No relationship was found between
maternal employment status and conﬂicts that females had with mothers or fathers.
Also, adolescents of both sexes spent less "free time" with their parents (both mother
and father) when their mothers worked, especially when they worked full time, than
when they were nonemployed.

In a review of the effects of mothers’ employment on marital relations, on
parent-child relations and on the mothers themselyes, Scarr, Philips and McCartney,
(1990), concluded that maternal employment per se was not the major issue in either

marital or parent-child relationships. Rather, the circumstances of the family, the

 

18

attitudes and expectations of fathers and mothers, and the distribution of time
available had important effects.

little is known about father involvement with school-aged children in different
family employment contexts. Crouter and Crowley (1990) examined father’s
involvement with their fourth and ﬁfth grade boys and girls in two family contexts:
dual-earner families in which both parents work outside the home, and single-earner
families in which the more traditional division of labor is practiced with fathers
employed outside the home and mothers engaged primarily in care of home and
children. Results indicated that fathers in single-earner families spent signiﬁcantly
more time in dyadic activities with sons than with daughters; dual-earner fathers
spent equal amounts of time with sons and daughters. Controlling for family size,
girls with greater dyadic involvement felt closer to their fathers than girls with less
dyadic involvement; boys feelings of closeness were not linked to levels of one-to—one

involvement with their fathers.

Income and Education

In reviewing the literature concerned with the effects of income and education
on parent-child relations, it is clear that these variables are usually considered as part
of a socioeconomic (SES) construct. Although there are different measures of SES,
it usually includes occupational status, income and education.

Gekas (1979) in reviewing literature in the area of the influence of social class

on family relations, reports the following generalizations from research:

19
. SES is positively related to equalitarian relationships between parent and
Child. SES is negatively related to autocratic relationships between parent and child.

SES is positively related to parental affection and involvement.

SES is negatively related to parental use of commands and imperatives.

SES is positively related to explanation of rules.

There is not an abundance of research in the area of parent-adolescent
relations in which socioeconomic status has been a consideration. The majority of
research has been conducted using middle and upper-middle class samples. Although
these studies frequently report no differences according to socioeconomic status
(SES), the sampling procedures do not adequately address populations of lower SES.

Although much of the research that has examined SES as an explanatory
variable in parent-child relations has focused on young children, that which is
concerned with parent-adolescent relationships suggests that parents’ socioeconomic
background may moderate the ways in which transformations in the parent-
adolescent relationship affect both parents and children (Jacob, 1974; Silverberg and
Steinberg, 1987).

Deﬁcient economic resources have been recognized as an important factor in
subjecting American families to excessive stress and children to developmental risk
(National Academy of Sciences, 1976).

Garbarino, Schellenbach and Sebes (1986) state:

Socioeconomics has a well-documented history as a marker variable

used as a basis for identifying differences in parenting styles, child-

rearing attitudes, parent-child relations and developmental outcomes
for children.

20

In Garbarino’s research on parental maladaptive behavior toward adolescents, SES
was found to be a key variable. Even when other variables were controlled, the
higher the SES the less maladaptive the parenting behavior. SES consisted of
maternal education, paternal occupational status and per capita family income.

Kohn (1979) deﬁned social class as "aggregates of individuals who occupy
broadly similar positions in a hierarchy of power, privilege, and prestige." The
principal components of social class in Kohn’s research were education, occupational
status and income, with the latter being of "distinctly secondary importance."
Education was found to be the more "potent" of the three components of SES (Kohn,
1969).

Education is an important variable to consider because of its relationship to
parenting. Education is an important determinant of occupational conditions and
powerfully related to parental values (Kohn, 1979). Education has a direct effect on
parental values, apart from its indirect effect mediated through job conditions;
education provides the intellectual ﬂexibility and breadth of perspective that are
essential for self-directed values. Garbarino et al. (1986) found an inverse
relationship between maternal education and maladaptive parenting; that is, the
higher the educational attainment of mothers, the lower was the rating of
maladaptive parenting.

With respect to income and parent-adolescent relations, an 1981 National
Center on Child Abuse and Neglect study found that 66 percent of the adolescent

maltreatment cases had family incomes below $15,000 (Garbarino, et al., 1986). The

21

adolescent maltreatment families were characterized by paternalistic, harsh, rigid
styles of child rearing and denial of feelings within the family. One can conclude that
these styles would be reﬂected in communication between parents and early
adolescents.

Although researchers generally combine variables to construct a measure of
SES, the distinct and separate effects of the variables such as income and education
can only be measured by keeping these variables singular and independent from one
another. In studies of the effects of maternal employment on children, it has been
shown that mediating factors should be considered (Bronfenbrenner and Crouter,
1982). These authors suggest that indices of social class should be differentiated in
order to determine effects speciﬁc to such aspects as each parent’s occupation,
education and contribution to family income. There is evidence that each of these
factors (income, education, and occupational status) qualiﬁes the impact of parental
work in a different way. We can logically conclude that these variables also have a
distinct, and in some cases differential, impact on other outcome variables as well,
such as parent-child relations.

More speciﬁc to communication within families, differences in self-disclosure
are noted (Beebe and Masterson, 1986) between working-class and middle-class
families. The research suggests that middle-class men tend to be higher disclosers
than lower-class men. While there are many exceptions, children tend to disclose
more to their mothers than to their fathers. Middle-class mothers tend to respond

verbally to their children more than lower-class mothers. Self-disclosure appears to

22

play a more pronounced role in the maintenance of family relationships among

middle-class families.

Palm-Job Role Stress

Research in the area of work-related stress and the stress involved in the joint
parental roles of employment and parenting is limited. Parent-job stress has been
found to affect the family (Piotrkowski, Rapoport and Rapoport, 1979; Mortimer,
Lorence and Kumka, 1986). It has also been found to exacerbate parent-child
conﬂict (Piotrkowski et al., 1987). However, parental work characteristics did not
explain any of the variance in emotional stress among early adolescents in a study of
dual-earner families (Bird and Kemerit, 1990). This study was limited to "earner" as
opposed to "career" families and included parents who were both employed full-time.

In a review of the literature vis-a-vis role stress experienced by mothers who
also pursue careers (Rapoport and Rapoport, 1978), there was evidence to document
the position that women experience a proliferation of roles rather than an orderly
transition from one role to another, and, therefore, are vulnerable to role stress (also
referred to as role strain). Role stress is higher for women than for men, and women
are less successful at compartmentalizing their roles than are men.

A "dual-career" family is a particular type of dual-earner family in which both
husband and wife pursue active careers and family lives (Rapoport and Rapoport,
1978). Dual-career families are distinct from dual-earner couples in their attitudinal

support of the equalization of power and domestic responsibilities, and in their belief

23

in career advancement. Some studies have shown, however, that in reality these
goals are not achieved and that women continue to be responsible for the home in
so-called "dual-career" families.

A study of dual-career families found that mothers engaged in careers
experience guilt and anxiety over perceived failures in mothering. Their husbands
approached the demands of parenting and career from a less emotional perspective
and thus, did not experience the role stress so prominent among the wives (Johnson
and Johnson, 1977).

There is no evidence that the role stress experienced by mothers in dual-
career families results in negative consequences for the children. In fact, there is
evidence that employed mothers increase efforts to makeup for the time they work
by devising activities directed toward enhancing the environment of the children.
(Johnson and Johnson, 1977). Also, these parents were more concerned about the
quality of their relationship with their children than with objective behavior. There
is also evidence that when children become adolescents, the role stress and anxiety
decreases for these working mothers (Hoffman, 1977).

The form that work-family role stress takes is likely to vary by social class
(Mortimer and London, 1984). For example, while the most important stress on a
single-parent, female-headed household is economic, this source of stress is less
important in a professional-managerial, single-provider, married-couple family.
However, that family is more likely to be challenged by male work absorption that

precludes desired family involvement. The alienating nature of work for blue-collar

24

men, as well as financial pressures, leave little energy for constructive family
involvement. Such conditions promote harsher discipline for children; wives are
isolated from men’s work roles. In blue-collar families, women remain responsible
for almost all domestic tasks. Potentially conﬂicting demands from two careers and
heavy demands for child rearing create pressures in dual-career families. In short,
different family employment patterns create different types of job-parent role stress
for husbands and wives. It would not appear, however, that one type of family is
freer from these stresses than another.

Galambos and Maggs (1990) examined mothers’ work-related and global
stress to determine the association between stress and changes in mother-adolescent
relations across a 6-month period and to determine whether mothers’ stress and
mother-adolescent relations predicted the adolescents’ psychosocial adjustment.
Global stress was related to the mother’s decreasing acceptance of the child and was
negatively related to adolescents’ psychosocial adjustment. However, mothers’ work-
related stress was not associated with mother-adolescent relations or adolescent
adjustment. These findings seem to conﬂict with a study (Piotrkowski, 1979) in which
parental job strain was found to negatively affect family relations. These authors
report no diﬁerences in the relationship between variables according to gender of
adolescent or socioeconomic status of the family.

The lack of research and sometimes conﬂicting findings in the area of parent-
job role stress and its affect on parent-adolescent relationships, indicates the need for

additional research in this area.

25
Paception of Family Stress by Mother and Father

In a review of social stress and the family, McCubbin and Patterson (1983)
deﬁne a stressor as a "life event or transition impacting upon the family unit which
produces, or has the potential of producing, change in the family social system. This
change may be in various areas of family life such as its boundaries, goals, patterns
of interaction, roles, or values." Family hardships are deﬁned as "those demands on
the family unit speciﬁcally associated with the stressor event." Both the stressor and
its hardships place demands on the family system which need to be managed. Family
stress (as distinct from stressor) is a state which arises from an actual or perceived
imbalance in the family’s functioning due to demands for adjustment caused by the
stressors and hardships. Stress becomes distress when it is deﬁned as unpleasant or
undesirable by the family unit.

The family’s ability to cope with stress depends not only on the actual stressor
and hardships experienced by the family, but also on the family’s resources for
dealing with stress (including ﬂexibility and communication skills), and the deﬁnition
that the family makes of the situation (McCubbin and Patterson, 1983). Thus, family
members’ perceptions of how difficult things are or have been is important in the
examination of the impact of stress of family interaction and coping. Moreover, the
family’s attempts to cope with stress result in a temporary disturbance in family
functioning. Role reorientation occurs with a revision of expectations accepted by

family members. Although family relationships will rebound in time, it can be

26

expected that family relationships may be perceived as more conﬂictual and perhaps
less open as a direct result of stress.

Psychosocial stress has been found to have a negative impact on family
relations. In a study of 785 Navy families, (Eastman, Archer and Ball, 1990), higher
life-stress families reported lower scores on cohesiveness, expressiveness and
organization as well as higher scores on family conﬂict. Economic stress in family
life has been linked to depression and demoralization for both parents--conditions
which were related to disruptions in skillful parenting (Conger, Conger, Elder and
Lorenz, 1992). In studying stress and relations between mothers and adolescents,
Galambos and Maggs (1990) found that global stress was related to the mother’s
decreasing acceptance of the child and negatively related to adolescent’s psychosocial
adjustment.

Family ecological theory (Bubolz and Sontag, 1990), assumes that processes
taking place between families and their environment, and well as within families,
must be viewed as interdependent. Thus, families are interdependent with their
environment and family members are interdependent with one another. The systems
approach to the study of family stress has shown the reciprocal nature of the impact
of stress in the family. In separate studies of families, with early adolescents as well
as teens--it was found that distress in some family members was signiﬁcantly
associated with stressors experienced by other family members, thus verifying that
distress is transmitted from one family member to another (Thomson and Vaus 1986;

Compas, Howell, Phares and Williams, 1989). These studies highlight the importance

27

of studying stress processes between individual family members. For example, the
amount of family stress perceived by the mother impacts the communication
relationship between mother and child in a reciprocal fashion. The perceived stress

of one family member has an impact on family relations as a whole.

Child is Perceived Acceptmzce by Motha and Father

Acceptance by the parent of the child is similar to parental warmth (Paulson
et al., 1991). Warmth is conceived as the extent to which a positive, benevolent
attitude permeates child rearing and is linked to self-disclosure and expression of
affection. Interpersonal acceptance has been labeled a "hallmark" of successful
family communication (Beebe and Masterson, 1986). Parental warmth has been
shown to be positively associated with self-esteem as well as expressiveness among
adolescents (Paulson et al., 1991).

The communication literature shows that close, interpersonal relationships are
built through self-disclosure, that is the sharing of information about oneself that
would not otherwise be known--revealing one’s inner world to another (Beebe and
Masterson, 1986; Cozby, 1973;). Moreover, the relationship between self-disclosure
and acceptance is a spiraling, reciprocal relationship. That is, the more a child feels
accepted by his parent, the more he discloses to that parent. The parent, in turn,
discloses more to the child, further building trust and intimacy, leading to increasing
levels of disclosure by the child. Clearly, acceptance is related to both negative and

positive communication behavior within the family. If a child feels rejected by

28

problematic communication between the parent and child, the parent-child relation-
ship will be negatively affected. Conversely, if the child feels accepted "no matter

what", communication openness is facilitated and the effects on the parent-child

relationship are positive.

Within Family D'qferences

A number of studies have found that adolescents have a more negative view
of family interaction patterns than parents, reﬂecting the adolescent’s need for
control and independence. Dalusio (1972), in examining self-disclosure between
parents and adolescents, found that parents reported more disclosure than their
children perceived. Barnes and Olson (1985), in comparing communication scores
within families, found that adolescents viewed their intrafamily communication with
greater negativism than did parents. Callan and N oller (1986) found perceptions of
interactions within the family were different for parents and adolescents. Family
members were rated as higher in anxiety, less involved and less dominant by
adolescents. However, with respect to friendliness of family interactions, adolescent
ratings were as high as those of parents. The authors’ conclusion from the ﬁndings
reﬂects the developmental paradox of adolescence: the perceptions of the adolescents
reﬂect their needfor autonomy as well as their desire for closeness within the family.

While research ﬁndings regarding differences between male and female
adolescents are not conclusive, it is clear that adolescents in general, have closer

relationships with their mothers than they do with their fathers. Noller and Bagi

29

(1985) found that both sons and daughters communicate more, and on a wider range
of topics, with mothers than fathers. Barnes and Olson (1985) found that mothers
reported better communication with their adolescents than did fathers. In terms of
a "target" for disclosure, that is, the person to whom disclosure is made, children and
adolescents in grades 4 through 12, regardless of gender, have been found to prefer
mothers over fathers (Rivenbark, 1971). A number of additional studies found that
adolescents report more disclosure to their mothers than to their fathers (Dalusio,
1972; J ourard, 1971; Komarovsky, 1974). Using an intimacy scale (which included
items dealing with empathy disclosure, consensus formation, and companionship),
Hunter and Youniss (1982) found that boys and girls reported greater intimacy with
their mothers than with their fathers in 4th grade, 10th grade, and in college. Offer
(1969) found that adolescent males viewed their fathers as more distant as less
understanding than their mothers. Adolescents from two-parent families were found
to have separate but coordinated relations with their mothers and fathers, with
fathers having a more instrumental role, while mothers are more accepting and open

(Smollar and Youniss, 1985 ).

CHAPTER 3

METHODOLOGY

THE DATA/ SAMPLE

The purpose of this study is to examine the inﬂuence of selected physiological,
sociological and psychological variables on communication between parents and early
adolescents. Information was obtained through structured in-home interviews of
early adolescents and their parents. The interviews were conducted as part of a
larger study, the Michigan Early Adolescent Survey (MEAS)II (1987), designed to
obtain information about a wide range of topics concerning youth and their families.

The sample design for the MEAS H was a cross-sectional survey of 245 youths,
ages ten to fourteen, and their parent or parents living in Michigan in 1987 (Keith,
H00pfer, Potter and Thompson, 1992). Each youth in the survey was personally
interviewed in his or her home by a trained volunteer. During the interview, which

lasted approximately one hour, parents (both mother and father, if present)

30

31

completed a questionnaire. Each family was assigned an identiﬁcation number in
order to facilitate analysis of the family as a system.

Interviewers were trained in the Spring of 1987 through intensive one-day
training sessions in four locations throughout the State of Michigan. The
interviewers were volunteers who were recruited by 4-H county program leaders and
program assistants in participating counties. During the training, interviewers were
given information regarding attitudes and issues of early adolescent children,
sensitized to nonverbal cues, given guidelines on the interview process, and provided
with the opportunity to critique videotaped interviews. The parent questionnaire was
also discussed.

Interviewers were similar in age to the parents of the early adolescents whom
they were interviewing; most were between the ages of 31 and 45. They were likely
to have had at least some college and to be employed.

Following interviewer training, letters explaining the purpose of the survey
were sent to each potential family (those randomly selected through the sampling
procedure). The interviewers contacted the families by telephone to determine

interest in being participants in the survey and to arrange an interview appointment.

SAMPLE SELECTION

The Michigan Early Adolescent Survey II sample included 245 cases drawn

from 24 Michigan counties using a stratiﬁed multistage cluster sampling technique.

32

This sampling procedure ensured that all Michigan early adolescents in public and
private schools who lived with their families had an equal chance of being chosen for
the study. The three stratum from which the sample was drawn represent the three
distinct geographical regions in the state: the southeast metropolitan region
(Macomb, Oakland and Wayne counties); southern Michigan (Barry, Barrien,
Calhoun, Cass, Genessee, Ingham, Jackson, Kent, Lapeer, Monroe, Montcalm,
Saginaw and Washtenaw counties); and northern Michigan and the Upper Peninsula
(Alpena, Baraga, Marquette, Mason, Montmorency, Ostego, Schoolcraft and
Wexford counties). Equal numbers of individual children and their parents were
selected from two school districts in each of these counties.

The sample used in this study was a subset of the larger sample of 245 early
adolescents and their parents who participated in the Michigan Early Adolescent
Survey H. The cases chosen were those that contained complete data for those
aspects of the survey pertaining to parent-child communication. The total number
of early adolescents who completed the communication portion of the interview was
224. Mothers who completed the communication portion of the survey numbered
194; communication data was available for 114 fathers. For purposes of multiple
regression analysis used in this study, only two-parent families with complete data for
the dependent and independent variables were included (n=74). The use of this
sample enabled a systems approach to data analysis with inclusion of family systems

variables (i.e., both mothers’ and fathers’ parent-job role stress) (see Chapter 4).

33
SAMPLE DESCRIPTION

Table 3-1 shows a comparison of the subsample (N=74) to the total sample
(N=224) on selected demographic variables. The ﬁndings of this study cannot be
generalized to all early adolescents and their families living in the state of Michigan
because of substitutions made during the sampling procedure and self-selection by
respondents. Compared to the 1980 State of Michigan census data proﬁle of early
adolescents and their families, the families in this sample--the larger sample as well
as the smaller subset (N-74)--lived in more rural areas and were less racially diverse
(Schlabach, 1989). Parents in this sample were somewhat older and more highly
educated than those in the State as a whole. However, the study is generalizable to

families of similar composition.

MEASUREMENT PROCEDURES

The major instruments used for data collection included: the Youth Interview
Schedule (to be used by interviewer when talking with the early adolescents); and the
Parent Questionnaires--one each for mother and father (to be completed by mother
and / or father, depending on family composition and availability of parents). Most
of the items in the interview schedule and parent questionnaire came from sources

used in other studies.

34

Table 3-1. Total Sample Compared to Multiple Regression Subsample.
Selected Demographics.

 

Multiple Regression

 

Total Sample
(N = 224) Subsample (N = 74)
Percent' Percent

Gender of Child

Boys 47 50

Girls 56 50
Age of Child

Ten 9 12

Eleven 21 17

Twelve 28 24

Thirteen 21 27

Fourteen 21 19
Ethnicity (Child)

White 92 96

Black 3 «-

Asian-American 2 «-

American Indian 1 ---

Other 2 4
Ethnicity (Father)

White 94 96

Spanish-American 1 ---

Black 4 1

Other 1 3
Education (Father)

8th Grade or less 2 1

Some High School 4 4

High School Graduate 24 28

Some College 35 31

College Graduate 21 22

Graduate or Prof. School 14 14
Source of Employment (Father)

Employed by Other 82 83

Self-employed 11 13

Unemployed 7 4

Table 3.1 Cont.

35

 

Total Sample Multiple Regression
(N = 224) Subsample (N = 74)
Percent' Percent
Employment Status (Father)
Full-time 96 96
Part-time 3 3
Temporary 2 2
Family Income
Less that $10,000 7 3
$10,001 - 20,000 14 12
$20,001 - 30,000 25 24
$30,001 - 55,000 42 46
$55,001 - 75,000 9 15
Over $75,001 2 «-
Residence
Farm 6 8
Rural Area 34 35
Small Town (Under 5,000 Pop.) 13 14
Town (5,000 - 25,000 Pop.) 21 26
City (25,000 - 100,000 Pop.) 9 4
City (Over 100,000 Pop.) 9 ---
Suburb of Large City 7 13
Ethnicity (Mother)
White 87 97
Mexican-American 1 1
Black 10 ---
American Indian 1 «-
Other 1 1
Education (Mother)
8th Grade or less --- «-
Some High School 5 3
High School Graduate 42 39
Some College 28 30
College Graduate 14 16
Graduate or Prof. School 12 12

36

Table 3.1 Cont.
Total Sample

Multiple Regression

 

(N = 224) Subsample (N = 74)
Percent' Percent

Source of Employment (Mother)

Employed by Other 53 74

Self-employed 8 8

Unemployed 3 1

Homemaker 35 15

Student 1 1
Employment Status (Mother)

Full-time 71 64

Part-time 24 33

Temporary 5 3

 

' Percent of those responding for each item, excluding missing data.

Parent(s) completed questionnaires while the early adolescents were being

interviewed (in a different room). Interviewers left the homes in which families were

interviewed with completed Parent Questionnaires and a completed Youth Interview

Schedule for the early adolescent. The interviewer left an additional seven-page

questionnaire with each early adolescent to be completed and mailed to the

researchers. The mail-in questionnaire is not relevant to this study.

INSTRUMENTS

The questions to determine the demographic/sociological variables were

developed for the purposes of the MEAS H survey. The question to determine

37
family income (see Table 3-2) was asked of both mother and father. Mothers’ and

fathers’ responses were highly correlated (r= .834); therefore, mothers’ responses
were used for the purposes of this study.

Physiological variables included gender of child and pubertal age. Gender was
determined by the interviewer. Pubertal age (PUBRTAGE) was determined by the
mother’s perception of her child’s pubertal age. The index used to assess mothers’
perceptions of pubertal age was adapted from Tanner’s work (Nelson, 1985). This
index is an attempt to locate early adolescent females and males on a developmental
continuum based upon physical body changes as rated by parents (see Table 3-2).
A continuum as opposed to an absolute number was used; since some responses
were between whole numbers, (i.e., a mark on a line between numbers) the values
(1 to 5) were multiplied by 10 so that responses between numbers could be given
whole number values. Fathers were also asked to assess their child’s pubertal age.
Mothers’ and fathers’ assessments of their children were highly correlated (r=.794);
thus, mothers’ responses were used to determine the pubertal age of the early
adolescent.

The psychological variables were, for the most part, adapted from other
sources. See Table 3-2 for operational deﬁnitions, questions as stated and items
included in scale formations. The scales that were included in their entirety, as
originally developed, were those measuring the eight dependent (endogenous)

variables of parent-child communication. All eight of the endogenous variables were

38

Table 3-2. Overview of Item and Scale Formulations by Variable.

 

 

 

 

Gender of child (GENDER)

Mothers’ assessment of child’s
pubertal age (PUBRTAGE)

Male/female; coded as dummy
variable:

0 = female
1 = male
As measured by scale based on

Tanner (Nelson, 1985); scored 10 to
50, less mature to more mature.

 

Coded by interviewer on page 1 of child
questionnaire.

Puberty, or becoming an adult physically,
takes several years; it may begin when a
child is 10 or it may not begin until a
child is 16. There are some signs parents
can notice to indicate where a child may
be in this time of change. Think about a

S-point scale:

1 = physically a child
2 = slight changes physically
3 = several noticeable physical

changes

4 = many changes
5 = physically an adult

Place an X on the line near the number
which you think best describes your child.
Words written below to give examples of

the physical changes you may have
observed.
PHYSICALLY
CHILD

M 1 m

|

l
Slight breast |
changes 2 Slight dev.
Beginning body I body hair
hair I

|

l

| Some body hair,
Some breast dev. I upper lip, chin,

3 etc.
Height spurt | Some voice
Some body hair I change

I

I
Period begun 4 Height spurt

| 3 to 6 in.

l

l

| Shaving

| Muscles well-
Adult height I developed
Full breast dev. 5 Voice changed

PHYSICALLY

ADULT

Table 3-2. (Cont)

  
 

 
  

Independent Variables

Mother’s parent-job role stress
(MWKSI’RES)

Father's parent-job role stress
(DWKSI‘RES)

Child’s perception of acceptance by
mother (MACCEPT)

Child’s perception of acceptance by
father (DACCEPT)

Mother’s perception of family stress
(MFAMSTRS)

Father’s perception of family stress
(DFAMSI'RS)

   

Independent Variable:
Family income (INCOME)

Mother’s education (MOMEDUC)

39

Operational Definition

 

   

As assessed by mother‘s response to
one question, four point intensity
scale.

Same as above with father as
respondent.

As assessed by child’s response to
one question, 5 point intensity.

See above.

As assessed by mother’s response to
one question, four point intensity.

Same as above with father as
respondent.

As determined by mother’s response
to 6-point scale, low to high income
categories; coded as dummy variable,
0 = 1 and 2

= 3 through 6.

As determined by mother’s response
to 6-point scale, low to high
education categories.

 
 

To what extent are both being employed
and being a parent stressful to you?
Circle one number.

1 =- Not much at all

2 - Not much

3 - Somewhat
4 = Very much

Same as above.

How much does your mother accept you
no matter what your do?

1 = Not at all

2 = A little

3 = Some

4 = A lot

5 -= Very much

Same as above with father as referent.

Families often have good times and bad

times. Think about your family in the

past year or two. Circle the number of

the words that best describe how you feel

about your family.

1 = lthasbeenaverygoodtime

2 = lthasbeenbothgoodandbad,
mostly good

3 = It has been a difﬁcult time

4 = It has been a very difficult time

Same as above.

Please circle the amount that comes
closest to your total net income before
taxes last year in 1986 (include all forms
of income).

1. Less than $10,000

2. $10,001 - 20,000

3. $20,001 - 30,000

4. $30,001 - $5,000

5. $55,001 - 75,000

6. Over $75,000

1. 8th grade or less

2. Some high school

3. High school graduate

4. Some college

5. College graduate

6. Graduate or professional school

Table 3-2. (Cont)

Father's education

Parental employment status

TWO-parent family employment

typology as follows:

1) 2 full-time employed parents (FF);

2) 1 employed full-time, 1 employed
part-time (FP) or temporary;

3) l employed full-time, l
homemaker (FN)

Dependent Variables

Child’s rating of father (openness)
(CDADO)

40

Operational Definition

As determined by father’s response
to 6-point scale.

As determined by mother’s and
father’s responses to type and time
spent in employment; coded as
dummy variable,

0 - FP

1 = FF and FN

Child’s response to lO-item openness
subscale of Olson’s Parent-
Adolescent Communication Scale; 1-5
likert scale; mean of 10 items, father
as referent.

Questions as Stated
See above.

Which best describes you major type of
employment (circle one).

1. Self-employed

2. Employed by other

3. Unemployed

4. Student

5. Homemaker

1. Full-time
2. Part—time
3. Temporary

This next section asks about how you talk
with your (child/mother/father). Circle
the number of the answer that best
describes your relationship with your
child.

1 a Strongly disagree

2 = Moderately disagree

3 = Neither agree nor disagree

4 = Moderately agree

5 = Strongly agree

° 1. I can discuss my beliefs with my
(child/mother/father) without
feeling restrained or embarrassed.

+ 2. Sometimes I have trouble believing
everything my (child/mother/father)
tells me.

° 3. My (child/mother/father) is always
a good listener.

+ 4. I am sometimes afraid to ask my
(child/mother/father) for what I
want.

i 5. My (child/mother/father) has a
tendency to say things to me which
would be better left unsaid.

° 6. My (child/mother/father) can tell
how I’m feeling without asking.

° 7. I am very satisﬁed with how my
(child/mother/father) and I talk
together.

‘ 8. If I were in trouble, I could tell my
(child/mother/father).

' 9. l openly show affection to my
(child/mother/fathcr).

+10. When we are having a problem, I
often give my (child/mother/father)
the silent treatment.

Table 3-2. (Cont)

Child’s rating of father (Cont)

Child’s rating of mother (openness)
(CMOMO)

Child’s rating of father (problem-
free) (CDADP)

Child’s rating of mother (problem-
free) (CMOMP)

Father’s rating of child (openness)
(DCl-lTLKO)

Mother’s rating of child (openness)
(MCHTLKO)

Father’s rating of child (problem-
CHT'LKP)

free) (D

Mother’s rating of child (problem-
free) (MCHTLKP)

41

Operational Definition Questions as Stated
4'11. I am careful about what I say to my
(child/mother/father).

2. When talking with my (child/
mother/father), l have a tendency
to say things that would be better
left unsaid.

. When I ask questions, I get honest
answers from my (child/mother/
father).

. My child tries to understand my
point of view.

+15. There are topics I avoid discussing
with my child.

. I ﬁnd it easy to discuss problems
with my child.

17. It is very easy for me to express all

my true feelings to my child.

‘18. My child nags/bothers me.

’19. My child insults me when s/he is

angry with me.

“‘20. I don’t think I can tell my child how

I really feel about some things.

4,

p—a

y-I
b)

p.‘
A

H
O\

' openness subscale
+ problem-free subscale (reverse-scored)

Same as above with mother as See above.

referent.

Child‘s response to lO—item problem-
free subscale (Olson); mean of 10
items; father as referent.

Same as above with mother as
referent.

Father’s response to 10-item
openness subscale (Olson); mean of
10 items; child as referent.

Same as above with mother as
respondent.

Father’s response to 10-item
problem-free subscale (Olson); mean
of 10 items; child as referent.

Same as above with mother as
respondent.

measured using the same instrument. The adolescents, as well as their parents, were

asked to complete the Parent-Adolescent Communication Scale (Barnes & Olson,

42

1982), a 20-item instrument composed of two subscales--one that measures the
degree of openness in communication, and one that assesses the extent of problems
in communication (see Table 3-2). Respondents (mother, father and early
adolescent) used a ﬁve-point Likert scale to indicate the extent of their agreement
with the items. The scores for items on the problems subscale are reversed in value.
Thus, a high score indicates a lack of perceived problems in communication
(problem-free). For purposes of analysis, the subscales (problem-free and openness)
are examined separately for mothers, fathers and adolescents (see analysis section).
The only difference between the parent and adolescent forms of the scale is the
referent; adolescents responded to the same items twice-once using their mother as
the referent and again with their father as the referent.

Missing data for the scales used in this study (Parent-Adolescent
Communication Scales) were handled using a ten percent rule; for scales with ten
items (i.e., openness and problem-free communication), when two items were
missing, the scale was not created for that person. If only one item was missing, the
response for that item was scored as the average of the other nine items in the scale.
The result of missing data on a few of the scale items resulted in a slight reduction
in the number of subjects in the analysis.

Cronbach’s reliability coefﬁcients for the dependent variables on each subscale

for this sample are as follows:

43

Openness Problem-free

Early adolescents’ rating of mother .69

Early adolescents’ rating of father .80 .73
Mothers’ rating of early adolescent .72 .75
Fathers’ rating of early adolescent .74 .67

These reliability coefficients indicate that responses of mother, father and
early adolescent to the 10 items on the problem-free and openness subscales are
sufficiently intercorrelated to demonstrate internal reliability. Alpha reliabilities for
other samples, using these same scales with older adolescents (age 12 to 20) have
been found to be slightly higher (Barnes and Olson, 1985); the coefficient for the
openness subscale was .87 and the problem-free scale was .78 (average for mother,
father and adolescent). A separate study showed test-retest reliability to be .78 and

.77 for the openness and problem-free subscales, respectively.

THE STATISTICAL MODEL

Multiple regression analysis was the statistical procedure used to estimate the
parent-child communication relationships. An overview of the variables included in
the model is shown in Table 3-3. A brief description of the exogenous (independent)
variables (grouped by ecological type) and the endogenous dependent variables is
shown. A system of eight multiple regression equations was used to examine
communication within the family system--mother, father and early adolescent. The

full specified eight equation statistical model is shown in Table 3-4. Alpha levels

44

were set at .05; however, precise p-values are reported (see Table 4-2, Chapter 4) for

each variable included in the ﬁnal regression model.

Table 3-3. Overview of Research.

 

Dependent Variables

Early adolescents’ rating of mother Mothers’ rating of early adolescent

(openness) (openness)
Early adolescents’ rating of mother Mothers’ rating of early adolescent
(problem-free) (problem-free)
Early adolescents’ rating of father Fathers’ rating of early adolescent
(openness) (openness)
Early adolescents’ rating of father Fathers’ rating of early adolescent
(problem-free) (problem-free)
Independent Variables
Physiological Variables Sociological Variables
Gender of early adolescent Family income
Pubertal age of early adolescent Education of mother

Education of father
Parental work status

Psychological Variables
Mam

Early adolescents’ perception of acceptance by mother
Early adolescents’ perception of acceptance by father

Parental Variables
Mothers’ job-parent role stress
Fathers’ job-parent role stress
Mothers’ perception of family stress
Fathers’ perception of family stress

Note: For specific regression model, identifying the independent and dependent variables for each of
the eight equations, see Table 3-4.

Table 3-4. Eight Equation Statistical Model of Parent-Child Communication.

 

1- Yl=ﬁo+ﬁlxl+82X2+33X;+AX,+ﬁ5X5+B6X6+B7X7+BEX8+£9X9+Bme+c

Dependent Variable:
Y1 = Child’s rating of father (openness)

Independent Variables:

X1 = Gender of child

X2 = Pubertal age of child

X3 = Family income

X, = Education of father

X5 = Job-parent role stress (mother)

X,s = Job-parent role stress (father)

X7 = Parent employment status (if)

X, = Parent employment status (fn)
(fp is the base)

X9 = Acceptance of child by father

X,o = Father’s perceived family stress

2- Y2=BO+BIXI +132X2+ﬁ3X3+ﬂuXH +55X5+56X6+ﬂvx7+ﬁsxs+ﬁizx12+313xis+c

Dependent Variable:
Y2 = Child’s rating of mother (openness)

Independent Variables:

X1 = Gender of child

X = Pubertal age of child

X3 = Family income

Xll = Education of mother

X5 = Job-parent rolc stress (mother)

X6 = Job-parent role stress (father)

X7 = Parent employment status (ff)

X,3 = Parent employment status (fn)
(fp is the base)

Xl2 = Acceptance of child by mother

X13 = Mother’s perceived family stress

N

3- Y3=Bo+51x1+52x2+33x3+54x4+Bsx5+ﬁsx¢+ﬁ7x7+ﬁsxs+ﬁgxg+ﬁwxm+c

Dependent Variable:
Y3 = Child’s rating of father (problem-free)

Independent Variables:

= Gender of child
X2 = Pubertal age of child

= Family income
X4 = Education of father
X5 = Job-parent role stress (mother)
X5 = Job-parent role stress (father)
X7 = Parent employment status (ff)
X8 = Parent employment status (fn)

(fp is the base)

X9 = Acceptance of child by father
X“, = Father’s perceived family stress

Table 3-4 (Cont)

4- Y4=ﬁo+ lei + 32X2+B3X3 + 511Xn+ 35x5 +ﬁ6)(6+ B,X7+ ﬁaxs'l' 5,2X,2+ﬂ,3X,3 + c

Dependent Variable:
Y, = Child’s rating of mother (problem-free)

Independent Variables:

X, = Gender of child

X2 = Pubertal age of child

X3 = Family income

X,, = Education of mother

X5 = Job-parent role stress (mother)

X,S = Job-parent role stress (father)

X7 = Parent employment status (fl)

X8 = Parent employment status (fn)
(fp is the base)

X,2 = Acceptance of child by mother

X,3 = Mother’s perceived family stress

5- Y5=ﬂo+ﬂlx1+BZX2+BBXB+B4XA+BSXS+B6XG+B7X7+BEXS+B9X9+BIOXIO+e

Dependent Variable:
Y5 = Father’s rating of child (openness)

Independent Variables:

X, = Gender of child

X2 = Pubertal age of child

X3 = Family income

X, = Education of father

X5 = Job-parent role stress (mother)

X,, = Job-parent role stress (father)

X7 = Parent employment status (ff)

X,3 = Parent employment status (fn)
(fp is the base)

X9 = Acceptance of child by father

X,o = Father’s perceived family stress

6- Y6: 130+ ﬁlxl + ’32X2 + 53x3 +3uxn + ﬁsxs + Boxa‘l' 37x7+ﬁsxs+ B,2X,2+ 513x13 + e

Dependent Variable:
Y6 = Mother’s rating of child (openness)

Independent Variables:

X, = Gender of child

X2 = Pubertal age of child

X3 = Family income

X,, = Education of mother

X5 = Job-parent role stress (mother)

X,, = Job-parent role stress (father)

X7 = Parent employment status (ff)

X,, = Parent employment status (fn)
(fp is the base)

X,2 = Acceptance of child by mother

X,3 = Mother’s perceived family stress

Table 3-4 (Cont)

7° Y7=Bo+5rxi+32x2+3sxs+54x4+ﬂsxs'iﬁais'l'ﬁxa‘wxs'l‘ﬁsoxol'ﬁtoxiol'c

Dependent Variable:
Y7 = Father’s rating of child (openness)

Independent Variables:

X, = Gender of child

X2 = Pubertal age of child

X, = Family income

X4 = Education of father

X5 = Job-parent role stress (mother)

X,5 = Job—parent role stress (father)

X7 = Parent employment status (ff)

X,, = Parent employment status (fn)
(fp is the base)

X9 = Acceptance of child by father

X,,, = Father’s perceived family stress

8° Y8=B0+B,X,+BZX2+£3X3+B,,X,,+35X5+36X6+B,X7+B,,X8+B,2X,2+ﬁ,3x,3+e

Dependent Variable:
Y8 = Mother’s rating of child (openness)

Independent Variables:

X, = Gender of child

X2 = Pubertal age of child

X3 = Family income

X,, = Education of mother

X, = J ob-parent role stress (mother)

X,, = J ob-parent role stress (father)

X7 = Parent employment status (ff)

X, = Parent employment status (fn)
(fp is the base)

X,2 = Acceptance of child by mother

X,3 = Mother’s perceived family stress

HYPOTHESIZED RELATIONSHIPS

Based upon the developmental contextual view of human development as

described in Chapter 1 and the review of literature as summarized in Chapter 2,

twelve hypotheses are stated. They are grouped according to environmental context:

physiological, psychological, and sociological. In addition, the strength of the

relationship of communication ratings within the family--mother, father and early

48

adolescent--were examined using Pearson correlation coefficients; expectations

regarding within family correlations are delineated.

Physiological

H1: There is a signiﬁcant negative relationship between pubertal age and
early adolescents’ ratings of their mothers and fathers as well as mothers’ and fathers’
ratings of their early adolescents (openness and problem-free).

H2: The effect of early adolescent gender on each of the dependent variables

is not signiﬁcantly different from 0.

Psychological

H3: There is a signiﬁcant negative relationship between maternal parent-job
role stress and mothers’ ratings of early adolescents as well as the early adolescents’
ratings of their mothers (openness and problem-free).

H4: The effect of paternal parent-job role stress on each of the dependent
measures is not significantly different from 0.

H5: There is a significant positive relationship between the early adolescents’
perceptions of acceptance by their mother and their ratings of their mother as well
as mothers’ ratings of early adolescents (openness and problem-free).

H6: There is a significant positive relationship between the early adolescents’
perceptions of acceptance by their father and their ratings of their father as well as

fathers’ ratings of early adolescents (openness and problem-free).

49

H7: There is a signiﬁcant negative relationship between mothers’ perceptions
of family stress and mothers’ ratings of their early adolescents as well as early
adolescents’ ratings of their mothers (openness and problem-free).

H8: There is a signiﬁcant negative relationship between fathers’ perceptions
of family stress and fathers’ ratings of their early adolescents as well as early

adolescents’ ratings of their fathers (openness and problem-free).

Sodological

H9: There is a significant positive relationship between family income and
mothers’ and fathers’ ratings of early adolescents as well as early adolescents’ ratings
of mothers and fathers (openness and problem-free).

H10: There is a significant positive relationship between the education of the
mother and mothers’ ratings of early adolescents as well as the early adolescents’
ratings of their mothers (openness and problem-free).

H11: There is a signiﬁcant positive relationship between the education of the
father and fathers’ ratings of early adolescents as well as the early adolescents’
ratings of their fathers (openness and problem-free).

H12: The effect of parental employment status on each of the dependent
variables is not significantly different from 0.

The above hypotheses will be tested through the application of the model
presented in Table 3-4. In addition, the strength of the relationship of

communication ratings within the family--mother, father, early adolescent--are

50

examined using Pearson correlation coefﬁcients. It is expected that early adolescents’
ratings of mother and father (openness and problem-free) will not be signiﬁcantly
correlated with parental ratings. It is also expected that early adolescents’ ratings of
mother will not be significantly correlated with early adolescents’ ratings of father
(openness and problem-free). Lastly, a significant positive correlation is expected
between mothers’ and fathers’ ratings of their early adolescent (openness and

problem-free).

OVERVIEW OF STATISTICAL METHODS

Ordinary least squares (OLS) regression was used to determine the relative
inﬂuence of each of the independent variables on the dependent variables. Multiple
regression analysis is a highly general and therefore very ﬂexible system for data
analysis that may be used whenever a quantitative variable (the dependent variable)
is to be studied as a function of, or in relationship to, any factors of interest
(expressed as independent variables). The form of the relationship between the
independent and dependent variables is not constrained; it may be simple or
complex, e.g., linear line or curvilinear (Cohen and Cohen, 1975).

Speciﬁcally, forced entry multiple regression was chosen as the most
appropriate tool for purposes of this study. This procedure has greater validity than
stepwise multiple regression in which there can be very serious capitalization on

chance (Cohen and Cohen, 1975). In stepwise regression the signiﬁcance test of the

51
contribution of an independent variable to multiple R-square proceeds in ignorance
of the large number of other such tests being performed at the same time for the
other competing variables; thus, neither the statistical signiﬁcance tests for each
variable nor the overall tests on the multiple R-square at each step are valid.
Lewis and Beck (1983) point out that in order to make accurate inferences
about the actual population parameter values, OLS regression must meet the
following assumptions.
1. No speciﬁcation error.
a. No relevant independent variables have been excluded.
b. No irrelevant independent variables have been included.

c. Correct functional form

2. No measurement error.
a. The variables X and Y are accurately measured.

3. The following assumptions concern the error term:

a. Zero mean: For each observation, the expected value of the error
term is zero.

b. Homoskedacsticity: The variance of the error term is constant for
all values of X.

c. No autocorrelation: The error terms are uncorrelated.

d. The independent variable is uncorrelated with the error term.

e. The error term is normally distributed.

Fortunately, assumptions are more likely to be violated with time series data
than with cross sectional data as in this study. Cohen and Cohen (1975) point out
that even a fairly substantial violation of the assumptions will frequently result in
little error of inference.

Although ordinary least squares regression was ultimately used for the

analysis, a number of statistical tests and evaluations were performed in order to

judge the best model for the purpose at hand. These included testing for

 

52

simultaneity and the existence of signiﬁcant cross-equation correlations. For a
description of the tests performed and alternative estimation procedures considered,
see Chapter 4.

SPSS-X was the primary statistical package used for data analyses. A second
package, SHAZAM, was also employed for testing and procedures not available on
SPSS, including testing for simultaneity and the existence of cross-equation

correlations.

LIMITATIONS OF THE STUDY

Survey research, utilizing self-report observations, does not always result in
accuracy or validity with respect to human behavior. Although a family member’s
perception is a measure of what he or she believes to be true, it may or may not
coincide with what an independent observer would ﬁnd to be true. Families are
often not their own best informants. However, gathering data from several family
members, as in this study, is superior to studies which utilize responses from one or
two members.

In addition, this study is a cross-sectional as opposed to a longitudinal study;
that is, data were collected at one point of time. In order to fully study the dynamic,
reciprocal nature of communication processes within the family and the impacts of
environmental factors on family communication patterns, changes over time should

be considered.

53

Limitations due to sampling are also evident. The procedure used resulted
in self-selection by families, depending on whether or not they decided to participate
in the study. As compared to the State of Michigan as a whole, this sample was
more highly educated and had higher family incomes; thus, the study is generalizable
only to families with similar demographic characteristics. Autocratic parenting styles,
inﬂexibility and the use of commands as opposed to inductive reasoning have been
linked to lower socioeconomic families. Thus, interpretations and responses on the
Parent-Adolescent Communication Scale may be very different for these families
than for those of lower socioeconomic
status. The limited number of low income families in this sample did not enable an
adequate assessment of differences for these families.

Finally, the use of multiple regression poses some limitations for this study.
Multiple regression assumes a certain structure among response categories; that is,
responses are assumed to be continuous measures with equal distances between
categories. Although this is not a completely accurate assumption--particularly with
respect to the independent variables in this study, imposing this am restriction
enables the application of a parsimonious regression model. Alternative estimation
procedures would likely require more parameters to be estimated and therefore,
require additional data. Although analysis variance can be used to test for
differences among means, regression analysis reveals the additional information

about the nature and extent of the relationship as reﬂected by the estimated

 

 

54

regression coefﬁcients. Regression results merely show correlation among

relationships and not causality.

 

 

CHAPTER 4

ANALYSIS AND FINDINGS

OVERVIEW OF ANALYSIS

The overall purpose of the research is to examine the relationship between
parent-child communication and selected biophysical and psychosociological variables
within the family system. In order to test the twelve hypotheses presented in Chapter
3, ordinary least squares (OLS) regression was employed to examine the relative
importance and direction of inﬂuence of the independent variables on the parental
and early adolescent communication ratings.

The sample used for the eight equation model is a subset of the sample of 245
early adolescents included in the Michigan Early Adolescent Survey II. It consists
of 74 families in which the early adolescent lives with both mother and father. If
data were missing for any variable under consideration, that family was excluded
from the analysis; thus, the final count is 74 families. Since this part of the study

focused on the family as a system and included system variables such as family stress

55

56

and the job-parent role stress of both parents, it was necessary to exclude families in
which data were missing for mother or father. Single parents were excluded from
this analyses because in all but two cases, data from fathers were missing for these

families.

EXAMINTNG THE DATA FOR LINEARITY/MULTICOLLINEARITY

It was initially assumed that the data in this study were linearly related.
Although a negative relationship between pubertal age and the each of the eight
dependent variables was hypothesized, studies with older youth (Steinberg, 1971;
Rivenbark, 1981) have found a curvilinear relationship between pubertal age and
parent-child relations. Thus, it was important to examine the nature (shape) of the
relationship between pubertal age and parent-child communication. In plotting
pubertal age with the eight dependent variables, patterns of curvilinearity were not
observed. In order to investigate the possibility of curvilinearity with respect to other
variables as well as pubertal age, a logarithmic transformation was performed on all
the variables (with the exception of the three dummy variables) and the model was
re-estimated. The results did not show a better ﬁt between the transformed
variables. Thus, it was concluded that the relationship between the independent and
dependent variables can be appropriately described as linear.

When two independent variables, or a linear combination of independent

variables are highly correlated, the relative effects of an individual variable can be

57

obscured or masked. Thus, it is important to test for the existence of
multicollinearity. A correlation matrix of dependent and independent variables in
shown in Table 4-1. Although high correlation between pairs of data suggest
multicollinearity, low correlations do not ensure the absence of multicollinearity.
Linear combinations of the variables may exhibit high intercorrelations. In such a
case, high standard errors of the estimated coefficients relative to the size of the
coefﬁcient (low t-ratios) is an indication of high multicollinearity. The individual t-
ratios were examined at each stage of the estimation. A number of variables
included in the original model (see Table 3-4) were eliminated in the ﬁnal regression
results (reported in Table 4-2) because of low t-ratios suggesting high standard errors

of the coefﬁcients.

TEST RESULTS FOR CROSS-EQUATION CORRELATIONS
AND SIMULTANEITY

The following section outlines the steps taken to examine the use of ordinary
least squares multiple regression as the most appropriate type of least squares
regression for this study.

Due to the fact that the eight equations in the model (Table 3-4) are related
to one another and can be viewed as a system, the use of seemingly unrelated
regression (SUR) was explored. A seemingly unrelated model consists of a system

of equations that are considered as a group because they bear a close conceptual

58

 

 

_8.- 8.. . Jana 3:3 as 688 .o z
83.- 88. 88. 8a.. woNHr 38. 83. :88. :83. HEM—QUAR—
aom-Hc 8. .38. 88.- 88.. 83. 88. .88. 83. 88552
89.- ..39? 8.- Suﬁ.- vmoar 83. 88. 38. 33.- 8.83252
88.- 2.3.. m8Nr 8.2.- 33..- 89. 28.- 8. 82.. 9.8253
83.- 98.. N8.- 83. «DH.- 88: m8: «9:.- 8: wag—.8
88.- 38.- 83.- 83 .88.- 8.- 8.- $8.- 38.- may—PEG
88. .38. ammo. .85. AFN. 98. 83. SS: 98. DOA—man
58.- 9.8. 8.- 8. mum“. $9. 39. $8. 88. 025202
84 5.8. 8. 88.- . 88. Sam.- 33: .28.- .NQ-N; MUSE—E:
84 Knew. :33. .83. mug. 88. $9. $8. 8802
8A 83. .88. .83. man. .83”. 9.3. 0:802
84 :83. 88. :8. 330.- 8H. 5.80:
88; KS. ammo. 85. 34:. Cushion
84 :88. :08! 33. mEOEU
8A :88. 8mm. man—U
88A :38. ©2020
84 03:0
HUS—RE: 53:52 DEEDS mush—*0: OMEGA— LSOEU man—U 0:520 Gan—U “Beau—Eco

 

605355 =ezaoE=EEeO EEO-2.8.5 3:32.:er 5:22.80 523m .H-v ~35.

59

88A

ham—00¢G

:88. mph.
884 8a.
884

:8.
83.

:vwﬁ.

8A

EH00; gangs—2 gums—<5:

88.-

89.-

.38.
hNNH.
884

way—Pm; mahgn 0DGHQ<G UDQHEOE

88.-
$8.
N8a.
85.

.88.

88A

8.. .

89.-
33.-
Sea.-
8H.-
38.

38.-
884

0.55% 8:3;

WNNH.
38.-
mi“.-
88.
Q8.-
85.-
:vomv.
88A

3 638 we 2

:m00<n
ham—00¢:
gab—age:
mgr—ME“:—

WERE—z

make—39
Qua—man
Qua—MEGS
MUS—baa:
88:02
OM50:
88:09
OM50:
QED—£0
mA—dA—0
02020
Dan—0

"macaw—2.50

.250 TV 035.

60

Table 4-2. Items Measuring Parent-Child Communication Openness:
Percent of Mothers, Fathers and Children Agreeing with Each Statement.’

 

 

 

 

Children about Parents about
Parents Children

Items Mothers Fathers Mothers Fathers

Total Sub Total Sub Total Sub Total Sub
I can discuss my beliefs with my
parents (child) without feeling
restrained or embarrassed. 60 66 58 65 88 89 87 88
My parent (child) is always a good
listener. 70 81 61 64 52 57 47 49
My parent (child) can tell how I
feel without asking. 53 58 36 46 84 82 69 69
I am very satisﬁed with how my
parent (child) and I talk together. 67 70 66 67 84 81 70 70
If I were in trouble, I could tell my
parent (child). 68 70 54 55 90 8’7 78 77
I openly show affection to my
parent (child). 59 73 51 58 96 95 82 82
When I ask questions, I get honest
answers from my parent (child). 74 77 70 77 78 73 78 78
My parent (child) tries to
understand my point of view. 77 87 70 76 73 72 65 68
I ﬁnd it easy to discuss problems
with my parent (child). 58 66 47 45 81 80 68 64
It is very easy for me to express all
my true feelings to my parent
(child). 50 50 45 46 77 77 57 51

 

 

1 ' Moderately or strongly agree response categories were combined to obtain percentages.
2 Total sample, N = 224 early adolescents, 194 mothers, 114 fathers; total subsample, N = 74 early adolescents, mothers, and

fathers.

61

Table 4-3. Items Measuring Parent-Child Communication Problems:
Percent of Mothers, Fathers and Children Agreeing with Each Statement.‘

 

 

Children about Parents about
Parents Children
Items Mothers Fathers Mothers Fathers

 

Total Sub Total Sub Total Sub Total Sub

Sometimes I have trouble believing
everything my parent (child) tells
me. 38 38 37 28 46 43 38 39

I am sometimes afraid to ask my
parents (child) what I want. 40 41 41 39 19 24 20 20

My parent (child) has a tendency
to say things to me which would
be better left unsaid. 36 33 31 27 30 30 26 26

When we are having a problem, I
often give my parent (child) the
silent treatment. 42 46 35 35 22 26 15 18

I am careful about what I say to
my parent (child). 61 61 65 66 55 43 61 62

When talking with my parent
(child), I have a tendency to say
things that would be better left 42 42 36 39 36 37 29 28
unsaid.

There are topics I avoid discussing
with my parent (child). 56 60 56 64 31 24 40 42

My parent (child) insults me when
he/she is mad at me. 17 12 14 11 3O 20 24 24

My parent (child) nags/bothers
me. 27 23 18 22 24 32 16 11

I don’t think I can tell my parent
(child) how I really feel about
some things. 39 35 38 34 29 28 24 24

 

1 ° Moderately or strongly agree response categories were combined to obtain percentages.
2 Total sample, N = 224 early adolescents, 194 mothers, 114 fathers; total subsample, N = 74 early adolescents, mothers, and
fathers.

62

relationship (Zellner, 1963). If the error terms in each equation are not related
(uncorrelated), then there is no relationship among the equations. In such a case,
ordinary least squares is appropriate. However, if the error terms among equations
are correlated, then the use of SUR will usually yield more efﬁcient estimates of the
coefﬁcients; that is, estimates in which there is minimum variance of the parameter
estimates. Statistical efﬁciency gains can be made by taking into account information
in the error terms by using a systems estimation procedure, SUR.

In order to determine if SUR was more appropriate than ordinary least
squares multiple regression, a Chi-square test for the possibility of cross-equation
correlations was performed. The null hypothesis is that all pairwise equation error
covariances equal zero. Since the calculated chi-square, 430, (df=28) is greater than
the critical chi-square of 41.3 (p =0.05), the hypothesis of zero cross-equation
correlation is rejected. This test suggests that it is appropriate to estimate the eight-
equation system taking into account the information contained in the error terms of
each equation.

According to Zellner (1963), ordinary least squares estimates may still be
judged best when small differences between OLS and SUR occur. SUR parameter
estimates are shown in Appendix A-l. Since SUR estimates are statistically more
efficient than OLS, the standard errors of the SUR parameter estimates are smaller
than those in Table 4-4. It should be noted that SUR estimates are sensitive to
speciﬁcation error. For instance, if one of the eight equations is misspecified, it will

affect the other seven equations in the SUR model. Thus, the SUR estimates for the

63

.2315: 380:6 .3. .0... .oEumo: .5 95.qu .358 m. as. on. .2: 26:3 2 :83.— a. 82: 5:3 25:905.? 93 can 33. .8.-2.0 5.. PS mos—win autono— oﬁ. .3

 

v; - aaamameo Sum .Em .88 u a: .88 u

8.8 u a 8.8 n 8 H88 - m
Sun A3: A85
898 - 8:8 + 28+ - 83862 x em
Rd u aged so sum .88 ”8.8 u Ema ”88 u
and u 8 H88 - a
5.8 $8
88.8 - 89% - 539:8 5 am
$.8- sesame Sam .sm .8.-o a: was n
88.8 n 8 8.8 u a 8.8 u .2
81$ 33 5.3
H88 - 33.8 + S88 - 95.562 8 em
as - cesium eo sum .Bw 3.8.8 - a: ”8.8 -
mad - 8 8.8 u
5.8 $3:
838 - ES. n 95:59 W am
So u ossaumm so Sam .2m 58 u .mmmm ”8.8 u «m
«88 u 8 8.8 u a
3.: see
328 + 8% - .5036 Zen

8.8 H 0385mm eo Sam .Em :8 - .wmwa ”8.8 u ”m
88.8 n 8 H88 - a

 

A36 ans
82.8 + 8n? n 5,30 m em
8.8 - age-as Sam .3 .8.-o $2 “8.8 u
«8.8-8 8.8-8 8.8-8
$3 Gnu 83
88.8 - 888 + 88+ u 02020 N am
8.8- oaaameo Sam .3 .88 - mad .78-
83 u a «88 u a 888 H88-
38 and ﬁns 83
3.3 - Ea - 8% ES - 0950 H em
mm mzﬁz mﬁmzﬁn mos—.82 $83.2 .8893 $28

 
 
 

 

 

.338: “Baa—em an»: baa—8.5 ”.252 33.385 =e=a3==EEeU 220.22.:— .v.v ~35.

 

.83 89a 8805.. «a... 3.. .9389: .Ho 9338.— ..oﬁmo a. .9». 05 .2: 26:2. 2 :83— u. 80... 5:3 Sermons? 8.. v... 38. $3.25 3.. 8.. 323... 3.598 05. .8
1.. no. .80 “an... a... no .30. on.» .ae-wbwsauoqﬁ 6.. .850 on» a»: «8 ob. ..a. . no 3.2.9.9..

.858 « .ouaezwo an $08 acaa a mu .Enr

.8.... u .. N8... - a
.88 .88
H88... + $8... - n 88.52 m ..cm
as... u a .8... u .. 8o... - .. a8... .- .. 8H... - a
.88 .58 .88 .8.: .38
H38... + $8... . onH... + 8H... - 88... - n .0...an 5 ..em
«a... u a 5... u a 8.8 u .. es... u a
.88 5.8 .38 .28
NHNH... - 8v; - 88... + 83... - u 93.....on e ..em
«8... u a 8.8 u a so... u a
.88 .38 .88
RS... + $8... - $3... - n 9350.. m ..m
n .9520 v em
8:. u ..
.NHHV
88... + u “.950 m ..em
88... n m 88 u 8 ms... - a
.88 .88 .38
Sam... - 88... - 8:... - - 02035 N em
.8... u a ﬁn... .- a
.88 88-8
memo - 88... - n 05.6 H ..m

 

 

 

.25 v-.. 2.3.

65

remaining seven appropriately speciﬁed equations will be inﬂuenced by the equation
that is misspecified. Although the Chi-square test suggested efﬁciency gains from the
use of SUR, it does not indicate how much or how the results will be affected.
Despite the gain in efﬁciency, a careful comparison between the DIS and SUR
results suggests little discernable difference. Therefore, notwithstanding the
conceptual advantage of SUR, since the two results show little empirical difference,
the 01.8 results are considered best for describing the parent-child communication
relationships.

It was also necessary to test the eight-equation system for simultaneity. An
assumption of regression is that the variables on the right side of the equation
(independent or exogenous) are independent of the error term. The existence of an
endogenous variable on the right hand side of the regression equation results in a
violation of one of the assumptions of OLS; that is, the endogenous explanatory
variable is not independent of the error term. If simultaneity exists, OLS will
generate biased parameter estimates. In the presence of simultaneity, an appropriate
alternative statistical estimation procedure two-stage least squares (Zellner, 1963).
In this study, the existence of more than one endogenous variable in each of the
eight equations necessitated testing for simultaneity. If the test conﬁrms its existence,
the model would have to be re-estimated using a more sophisticated estimation
procedure.

Two variables (MACCEPT and DACCEPT) that were classiﬁed as

independent variables were tested to conﬁrm that they were indeed independent

66

(exogenous) as opposed to dependent (exogenous) with respect to parent-child
communication. The assumption that acceptance would be simultaneously
determined with the dependent measures of communication (openness and problem-
free) is based on research showing the reciprocal, dyadic effect between trust and
intimacy and self-disclosure (Beebe and Masterson, 1986; Gilbert, 1986). If the child
trusts the parent (as determined by high levels of perceived acceptance), then the
child is more likely to self-disclose, that is, communicate openly with the parent.
Self-disclosure has a reciprocal effect; if the child is open with the parent, the parent
is likely to be open with the child and vice-a-versa. This two-way causal relationship
is an empirical question; as such, a statistical test for simultaneity was performed.
Speciﬁcally, the Hausman (1978) speciﬁcation test of exogeneity of the variables
designated as exogenous was performed; the null hypothesis is no correlation
between the exogenous variables and the error terms. The results of the test for
each equation in the model was that the null hypothesis could not be rejected. That
is, although we suspected simultaneity, the test results supported the existence of no
simultaneity and therefore, more sophisticated estimation procedures were not
required. Although not reported here, this test was further veriﬁed by examining the
model estimated by both two-stage (281.5) and three-stage (3SLS) least squares.
Two-stage least squares is a limited information estimation procedure while BSLS is
a full information procedure that allows for the possibility of cross-equation
correlations within a simultaneous equations framework. All test results showed no

need to use 281.8 or BSLS; OLS and SUR are sufficient.

 

’67
In sum, although testing was done for simultaneity, the existence of
simultaneity was not conﬁrmed. In addition, the eight-equation system was examined
for cross-equation correlation. Although testing conﬁrmed the existence of
correlated error terms, SUR estimations showed little discernable difference from

estimations using OLS; thus, it was determined that OLS was the most appropriate

estimation procedure.

DESCRIPTIVES

The Dependent Variables
Early Adolescents’ Ratings of Mom and Dad on Openness ( CM CM 0 and CDADO)
All eight dependent variables were measured using a 10-item subscale of
Olson’s 20-item parent-child communication scale, one for openness and another
problem-free communication (see Chapter 3). Table 4-2 shows the percentages of
adolescents moderately or strongly agreeing with each of the ten statements for
ratings of mothers and fathers on communication openness. Percentages are shown
for the total sample (224 early adolescents, 194 mothers and 114 fathers), as well as
the subsample used for purposes of multiple regression (N=74). Nearly two-thirds
of the youth reported being "very satisﬁed" with how they talk with both their mother
and father. Areas of strength for both parents were listening skills, honesty, and
understanding of the child’s point of view. In comparing adolescent ratings of

mothers and fathers, early adolescents were more likely to agree that their mothers

 

 

 

 

68

were open in communication. Whereas over half of the youth reported that their
mother could tell how they feel without asking, only one-third reported that their
father could do so. Youths were also less likely to tell their fathers if they were in
trouble or to discuss problems with their fathers than with their mothers.

This sample of early adolescents rated their mothers higher on communication
openness than they rated their fathers. The potential range of the scale is 1-5. The
actual range for the sample of 74 two-parent families in the multiple regression
analysis was from 1.70 to 5.00 for both mothers and fathers with a mean of 3.85 for

mothers and 3.62 for fathers.

Early Adolescents’ Ratings of Mom and Dad on Problem-free Communication
(CMOMP and CDADP)

Table 4-3 shows the percentages of adolescents moderately or strongly
agreeing with each of the ten statements for ratings of mothers and fathers on
communication problems. About 15 to 25 percent of the early adolescents in this
study reported that their parent (mother and father) "insults me when he/ she is mad
at me," or "mags/bothers me." Between one-half and two-thirds of the early
adolescents reported being careful about what they say to their parents.

The potential range of the scale for problem-free communication is the same
as for openness--1 to 5. The actual range for this sample of 74 two-parent families

was from 1.40 to 4.50 for fathers, and 1.00 to 4.50 for mothers. This sample of early

69

adolescents rated their fathers higher in problem-free communication than their
mothers; the mean score for fathers is 3.15 whereas the mean for mothers is 3.02.

Bear in mind that the mean for the problem-free scale was computed by reverse-
scoring the items. Thus, a higher mean score indicates that communication is more

problem-free than does a lower score.

Mothers’ and Fathers’ Ratings of Early Adolescents on Openness M CHT LKO and
DCHTLKO)

Table 4-2 shows the percentages of mothers and fathers moderately or strongly
agreeing with each of the ten statements for ratings of early adolescents on
communication openness. Items that were rated highly by both mothers and fathers
include: discussion of beliefs with the child; openly showing affection to the child;
honesty from child; and feeling that the parent(s) could tell the child if he/ she were
in trouble. More than three-fourths of the parents (mother and father) agreed with
statements in these areas. The potential range of the scale for problem-free
communication is 1 to 5. The actual range for this sample of two-parent families was
2.60 to 5.00 for fathers, and 2.30 to 5.0 for mothers. Mothers rated their
communication openness with their early adolescents higher than fathers rated their
relationship with the early adolescents; the mean score for mothers is 4.03 whereas

the mean for fathers is 3.80.

 

 

 

 

70

Mothers’ and Fathers’ Ratings of Early Adolescents on Problem-ﬂee Communication
(MCHTLIG’ and DCHTLKP)

Table 4-3 shows the percentage of parents (mothers and fathers) moderately
or strongly agreeing with each of the ten statements for ratings of early adolescents
on communication problems. Those items that the greatest number of mothers and
fathers agreed with (approximately one-third to one-half of the parents in this
sample) were: being careful about what he/she says to the child; and, sometimes
having trouble believing what the child tells him/her. The potential range of the
scale for problem-free communication is 1 to 5. The actual range for the sample of
74 two-parent families was 3.43 for mothers and 3.42 for fathers.

As with early adolescent ratings, the items measuring problems were reverse-scored

to determine means for the problem-free scale.

Independent Variables

Of the 74 families in each of the eight regression analyses, gender of child
(GENDER) was evenly divided; 37 included mother, father and an early adolescent
girl and 37 included mother, father and an early adolescent boy. The potential
range for pubertal age (PUBRTAGE) was 10 to 50. The actual range was 10 to 50
with a mean score (boys and girls combined) of 28.80.

Perception of acceptance (by mother) of the early adolescent (MACCEPT)
was measured on a scale of 1 to 5; the actual range for this sample was 2 to 5. The
mean score was 4.07. Of the early adolescents in this subsample, 76 percent reported

feeling "a lot" or "very much" acceptance by their mother. Perception of acceptance

71
(by father) of the early adolescent (DACCEPT) had a potential range of 1 to 5 and

an actual range of 1 to S. The mean score for DACCEPT was 3.99. Of the early
adolescents in this sample, 70 percent reported feeling "a lot" or "very much"
acceptance by their father.

Family income (INCOME) was coded as a dummy variable for use in the
regression analysis. The original range on this scale was 1 to 6, reﬂecting income
categories of: 1) less than $10,000, 2) $10,001-20,000, 3) $20,001-30,000, 4) $30,001
to 55,000, 5) $55,001 to $75,000 and 6) over $75,000. Categories 1 and 2 were
combined and coded 0 to include incomes of $20,000 and less. Categories 3 through
6 were combined and coded 1. This recoding was done to differentiate between high
and low income levels. Fifteen percent of the sample reported incomes of $20,000
or less; 85 percent reported incomes of more than $20,000.

About one-fourth of both mothers and fathers reported that it had been a
"somewhat difﬁcult" or "very difﬁcult" time for the family over the past year or two.
The large majority of parents (three-fourths) indicated that it had been a "good" or
"mostly good time." Mothers’ perceptions of family stress (MFAMSTRS) had a
potential range of 1 to 4 and an actual range of 1 to 4. The mean for MFAMSTRS
was 2.12. Fathers’ perceptions of family stress (DFAMSTRS) had a potential range
of 1 to 4 and an actual range of 1 to 4. The mean for DFAMSTRS was 2.14.

The mothers in the study reported greater job-parent role stress than did
fathers. Over 60 percent of the mothers in the study reported "somewhat or "very

much" when asked about the extent of stress in being both employed and a parent;

 

72

only about one—third of the fathers reported "somewhat" or "very much" stress. The
potential range for mothers’ job-parent role stress (MWKSTRS) was 1 to 4; the
actual range was also 1 to 4. The mean for MWKSTRS was 2.64. The potential
range for fathers’ job-parent role stress (DWKSTRS) was 1 to 4; the actual range was
also 1 to 4. The mean for DWKSTRS was 2.12.

The potential range for mothers’ education (MOMEDUC) was 1 to 6; the
actual range was 2 to 6. The mean for MOMEDUC was 3.96. The potential range
for fathers’ education (DADEDUC) was 1 to 6; the actual range was also 1 to 6.
The mean for DADEDUC was 4.08.

Parental work status was coded as a dummy variable and included three types:
1) both parents employed full-time (FF), one parent employed full-time and the
other parent employed part-time (PP), and 3) one parent employed full-time and the
other parent not employed outside of the home (FN). Over half of the parents in
this sample (53 percent) were both employed full-time. About one-fourth (27
percent) were in families in which the husband worked full—time and the wife worked
part-time. About one-ﬁfth (20 percent) of parents were categorized as FNuone

parent working full-time and the other not employed outside of the home.

Multiple Regression Results
Early Adolescents’ Ratings of Father and Mother on Openness (CDADO & CM OMO)

The ordinary least squares regression results are reported in Table 4-4. The

regression equation for the early adolescents’ rating of their fathers on openness

 

73

(equation 1) explained 34 percent of the variation in the dependent variable and the
F-statistic was found to be signiﬁcant at the .001 level. The equation for early
adolescents’ ratings of mothers (equation 2) explained 25 percent of the variation in
communication openness and was also signiﬁcant at the .001 level.

Clearly, the most signiﬁcant variable with respect to the early adolescents’
ratings of both mom and dad was the child’s perceived acceptance by that parent
(MACCEPT & DACCEPT). As hypothesized (H5 and H6), early adolescents were
more likely to be open with mom and dad if they feel that their parent will accept
them "no matter what." With respect to equations 1 and 2, both DACCEPT and
MACCEPT are positive and highly signiﬁcant. As shown in Figure 4-1, given a one
unit increase in DACCEPT, a 0.367 unit increase in CDADO can be expected.
Similarly, a one unit increase in MACCEPT will yield a 0.233 unit increase in
CMOMO. Father’s openness was more responsive (greater slope) to the early
adolescent’s feeling of acceptance than was mother’s openness.

A signiﬁcant negative relationship was found between pubertal age and the
early adolescent’s rating of both mother and father; the coefﬁcient for pubertal age
was signiﬁcant at the .002 level of probability for mothers and .01 for child’s rating
of their father. The mean for pubertal age for this sample, based on mothers’
ratings, is 28.8 (on a SO-pt. scale) indicating that these 10-14 year-olds are at the early
stages of puberty; thus, the relationship is a linear one as opposed to curvilinear
which might be true of later adolescents who are more physically mature (Steinberg

and Hill, 1978; Steinberg, 1987).

 

74

CDADO. CMOMO

 

4.2

 
 

omomo vs. mace. t
3.8 _. ............................................. .p ...................

   

3.3 _. .................................................................
3.4 q .................................................................
3.2 _ ..................................................................

3 edado vs. daocopt

   

2'8 ...................................................................

 

 

 

2-8 I I l l I I I
1 1 .5 2 2.5 3 3.5 4 4.5 5

DACCEPT. MACCEPT

Figure 4-1. The Relationship Between Early Adolescents’ Perceived
Acceptance by Mother and Father and Communication Openness.

Contrary to hypotheses (H3), mothers’ job-parent role stress (MWKSTRES),
that is, her perceived stress involved in combining parent and work roles, was not
related to the early adolescents’ ratings of their mothers or fathers on communication
openness. As hypothesized (H4), fathers’ job-parent role stress (DWKSTRES) was
not found to be a signiﬁcant contributing variable to early adolescents’ ratings of
their fathers (CDADO). Interestingly, however, paternal job-parent role stress
approached significance (p= .08) in children’s ratings of their mothers. This suggests
that for these children, the father’s situation regarding his dual job and parenting
roles, affects the mother-child relationship. Mother’s job-parent role stress
(MWKSTRES) was also negatively related to youth ratings of mother on openness
but was not signiﬁcant, indicating that father’s stress in combining these roles is a

greater factor in communication ratings than is mothers’ work stress.

 

 

75
The early adolescents did not rate their mothers (CMOMO) differently

depending on parental work status, i.e., two full-time employed parents (FF), one
full-time and one-part-time (FF), and one full-time and one not employed
(homemaker). However, the early adolescents’ rating of their fathers (CDADO)
appeared to be inﬂuenced by parental workstatus. Children in families in which
mothers were not employed outside of the home, (FN), rated fathers lower than in
families in which mothers were employed either part-time or full-time. A negative
relationship (p = .057) was found for families in which mothers were not employed
outside the home and adolescent ratings of fathers on openness. Perhaps in families
in which mother stays at home, she is the primary nurturer. Fathers in these families
may be cast in the role of breadwinners as opposed to nurturers; thus, children in
these families rate their fathers lower on openness than in families where mothers
work, reﬂecting the closer relationship between mothers and children in these
families. This ﬁnding does not support the hypothesis (H12) which states that work
status is not a contributing factor to differences in parent-child communication.

Gender was a contributing factor to the early adolescents’ ratings of mothers
(equation 2). Boys were signiﬁcantly (p = .028) more negative than girls in their
ratings of mothers on communication openness. This is the only equation where
gender was found to be statistically signiﬁcant.

In examining the early adolescents’ ratings of their fathers, father’s perception
of family stress (DFAMSTRS), based upon the question, "how good or bad things

have been going in the family over the past couple of years," was signiﬁcant (p=.02)

 

 

76

in explaining early adolescent’s ratings of father on openness. With respect to youths’
ratings of mothers on openness, mothers’ perceptions of family stress (MFAMSTRS)
was examined; however, this variable was found not to be signiﬁcant in explaining
early adolescent ratings of mother and omitted from the ﬁnal equation.

Based on statistical tests, education (mother and father) and family income
were consistently insigniﬁcant and eliminated from the ﬁnal regression equations for

early adolescents’ ratings of mother and father on openness.

Early Adolescents’ Ratings of Father and Mother-Problem-free (CDADP and CM 0MP)

The multiple regression equation for the early adolescents’ rating of fathers
(problem-free), shown in equation 3, Table 4-4, explained just 4 percent of the
variation in the dependent variable with an F-statistic p-value of .24 (equation 3,
Table 4-4). The equation for early adolescents’ rating of mothers (problem-free),
shown in equation 4, explained 4 percent of the variation in ratings with a
signiﬁcance level of 0.10. As with the child’s rating of mother and father on
openness, the child’s feeling about acceptance was most important; in fact, it is the
only variable that approached signiﬁcance in the examination of the early
adolescents’ ratings of mothers and fathers on problem-free communication with
signiﬁcance levels of 0.052 for mothers and .078 for fathers. With respect to early
adolescents’ ratings of their fathers, education of their father (t= 1.12) is positively

related to problem-free communication.

 

77
Mothers’ and Fathers’ Ratings of Early Adolescents-Openness (DCHTLKO and

MCHTLKO)

The regression equation for fathers’ ratings of the early adolescent on
openness (equation 5) explained 20 percent of the variation in the dependent
variable and was signﬁcant at the .004 level. The equation for mothers’ ratings of
the early adolescent on openness (equation 6) explained 25 percent of the variation
in ratings of the early adolescent and was signiﬁcant at the .003 level.

Family stress (MFAMSTRS and DFAMSTRS) was found to have a strong
negative relationship to mothers’ and fathers’ ratings of their early adolescents on
communication openness, with signiﬁcance levels of .01 for mothers (equation 6) and
.043 for fathers (equation 5).

Differences can be seen in comparing mothers and fathers in terms of the
variables found to be important in explaining differences in communication ratings
of their early adolescents (equation 5 and 6). The extent to which the early
adolescent feels accepted by his mother (MACCEPT) was the most important
identiﬁed variable in explaining differences in MCHTLKO (t=3.76, p=.001). Even
though this is a measure of the child’s (not the mother’s) perception, it was a
signiﬁcant contributing factor in the mothers’ ratings. The extent to which the early
adolescent feels accepted by her father (DACCEPT) was not statistically important
in explaining fathers’ ratings of early adolescents on communication openness

(DCHTLKO). This difference reﬂects the closeness of the mother-child relationship;

 

 

 

78

mothers are closer to their children than fathers are and therefore are more affected
by the attitudes and perceptions of the child.

Paternal job-parent role stress (DWKSTRES) was inversely related to fathers’
ratings of children on openness, and inversely related to mothers’ ratings as well.
Maternal job-parent role stress (MWKSTRES) was negatively related, but not
signiﬁcant in explaining the variation in fathers’ ratings; however, mothers’ job-parent
role stress was found to be positively related to mothers’ ratings, although not
signiﬁcant. These ﬁndings support the importance of a family systems approach to
the study of parent-chﬂd communication. Clearly, for this sample of relatively highly
educated two-parent families, father’s employment had a greater effect on parent-
child communication than did mother’s employment. Although a greater percentage
of mothers reported feeling job-parent role stress than fathers, the stress reported by
fathers had a greater effect on both mothers’ and fathers’ ratings of their early
adolescents on openness.

Parental work status was somewhat important as an inﬂuencing factor in
mothers’ ratings of their early adolescents on openness (MCHTLKO). In examining
the three parental employment types (FF, FN and PP) mothers in dual-earner
families (FF) and families in which the mother was not employed rated early
adolescents lower on openness than did mothers who were employed part-time; for

dual-earner families, this variable approached statistical signiﬁcance (p=.l3).

 

79
The effects of gender of child, pubertal age, and family income were all found

to be insigniﬁcant in explaining differences in parental ratings of early adolescents

on communication openness.

Fathers ’and Mothers ”Ratings of Early Adolescents--Problem-free (DCHTLKP and
MCHTLKP)

Recall that this scale is designed to measure problems in communication; it
was reverse-scored to be consistent with the openness scale. Thus, a higher score
reﬂects fewer problems in communication.

The regression equation for fathers’ ratings of early adolescents on problem-
ﬁ'ee communication (equation 7) explained 19 percent of the variation in DCHTLKP
and was signiﬁcant at the .03 level of probability. The equation for mothers’ratings
(equation 8) explained 23 percent of the variation in MCHTLKP and was signiﬁcant
at the .001 level of probability.

Findings for mothers and fathers were quite different with respect to the
factors found to be related to parental ratings of early adolescents on the problem-
free scale. The most signiﬁcant contributing factor to mothers’ ratings of children
on problem-free communication was that of family stress. Mothers’ perceptions of
"good and bad" times as a family over the past year or so (MFAMSTRS) was
inversely related to mothers’ ratings of problem-free communication behavior of

adolescents. However, fathers’ perceptions of family difﬁculties or stress

 

80

(DFAMSTRS) was not signiﬁcantly related to fathers’ ratings of early adolescents on
the problem-free scale (DCHTLKP).

As with fathers’ ratings of early adolescents on openness, the educational level
of fathers was positively (p = .006) related to fathers’ ratings of early adolescents on
openness; mothers’ education is not signiﬁcant in explaining mothers’ ratings. An
important factor inﬂuencing mothers’ ratings of early adolescents on openness was
that of perceived acceptance of the early adolescent by the mother (MACCEPT).
The extent to which the early adolescent feels "accepted no matter what" by his
mother was positively related to mothers’ rating of the early adolescent on
communication openness. The perceived acceptance of the early adolescent by father
(DACCEPT) was not statistically important to the father’s rating of the child
(DCHTLKP). Again, these differences for mother and father reﬂect the closeness
of the mother-child relationship.

Fathers’ job-parent role stress was negatively related to mothers’ rating of the
child (MCHTLKP) whereas mother’s job-parent role stress was positively related,
although insigniﬁcant, with respect to mothers’ rating of these early adolescents on
problem-free communication. Mothers’ job-role stress appears to be more important
in explaining variations in fathers’ ratings (DCHTLKP) than the fathers’ job-role
stress; both are negative.

Parental work status (FF, FN) is a signiﬁcant contributing factor for the
father’s rating of his early adolescent on problem-free communication with fathers

from dual-earner families (FF) rating early adolescents lower on problem-free

81

communication (that is, more problematic) than fathers ﬁ'om the other two
employment types (FN and FF). On the other hand, parental work status was not
important in explaining mothers’ ratings of early adolescent on problem-free
communication (MCHTLKP).

Gender of child, family income and pubertal age were not found to be
important contributors to differences in parental ratings of early adolescents on

problem-free communication and were omitted from both equations (DCHTLKP and

MCHTLKP).

SUMMARY OF FINDINGS

Nearly all of the hypotheses set forth in Chapter 3 were supported to some
extent by the empirical results with the exception of family income which was never
found to be important in explaining differences in communication ratings. As
hypothesized (H1), pubertal age was negatively related to the child’s rating of his
mother and father on openness. However, pubertal age was not important in six of
the eight equations; thus, in general, the hypothesis was not strongly supported.

For the most part, the hypothesis (H2) that gender of child would not be
related to communication ratings was supported; only in Equation 2 was gender 3
signiﬁcant contributing variable-girls rated mothers higher on openness than did

boys; this ﬁnding is consistent with other research (Noller and Callan, 1990)

82

indicating that females disclose more than males; thus, daughters perceived their
mothers more open than sons view their mothers.

The early adolescents’ perception of acceptance by father (DACCEPT) was
hypothesized (H5) to be positively related to fathers’ and early adolescents’ ratings
on both openness and problems. The hypotheses was supported for early
adolescents, but not for fathers. Hypothesis (H6) regarding early adolescents’ and
mothers’ ratings was supported; MACCEPT was signiﬁcant (or approaching
signiﬁcance) in explaining differences in communication ratings by youth of their
mothers as well as mothers’ ratings of their children. This finding reﬂects the
closeness of the mother-child relationship and supports the literature ﬁndings with
respect to the bidirectional inﬂuence of communication leading to establishing trust
relationships and high degrees of self-disclosure.

Hypotheses regarding parental perceptions of family stress (MFAMSTRS and
DFAMSTRS) (H7 and H8) were generally supported; in five of eight equations,
family stress was negative and statistically signiﬁcant in explaining differences in
communication ratings. Mothers’ ratings were more affected than fathers’ by family
stress; however, ratings of early adolescents of their mothers (both on openness and
problem-free communication) were not signiﬁcantly inﬂuenced by family stress
reported by the mother. Fathers’ reported stress was a signiﬁcant contributing factor
in the youths’ ratings of their father on communication openness. This ﬁnding

suggests that a father’s ability to relate to his early adolescent child may be more

 

83

affected by family stress a mother’s communication with her early adolescent
children.

The hypothesis (H4) that fathers’ parent-job work stress (DWKSTRS) would
not be a signiﬁcant explanatory variable in explaining communication differences was
generally not supported. In four of the eight equations, DWKSTRS was inversely
related to the dependent variable and statistically signiﬁcant. Differences in
communication between mothers and children were inﬂuenced by fathers’ parent-job
role stress more than by mothers’ perceived parent-job role stress.

In general the hypotheses (H3) that mothers’ parent-job role stress would be
negatively related to communication ratings was not supported. In only one of the
equations (fathers’ rating of early adolescents on problem-free communication) did
this variable approach signiﬁcance. It is possible that for families where mothers
experience parent-job role stress, fathers are more involved in parenting and
therefore, fathers in this type of family are more aware of problems and rate their
children lower in problem-free communication. Maternal parent-job role stress was
simply not important in the mothers’ or early adolescents’ ratings of mothers. In fact,
contrary to hypotheses, maternal parent-job role stress was found to be positively
related to mothers’ ratings of early adolescents, although not statistically signiﬁcant.
The lack of importance of maternal parent-job role stress suggests that employed
mothers are successfully handling this dual role. The importance of paternal parent-

job role stress, even with respect to the mother-child relationship, indicates that

84

fathers’ work environment inﬂuences family relationship to a greater extent than
mothers’ work situation.

In ﬁve of eight regression equations, work status was not a signiﬁcant
explanatory variable; thus, the hypothesis that the inﬂuence of work status on parent-
child communication would not be signiﬁcantly difference from zero was, for the
most part, supported. However, in three of the eight equations work status was a
signiﬁcant variable. In general, when compared to other employment types, families
in which fathers were employed full-time and mother part-time (FP) fared better
than those in which both parents were employed full-time or the mother was not
employed outside of the home; that is, FF and FN employment status had a negative
relationship to the dependent measures in comparison to FP families.

Based on the literature with respect to socioeconomic status, a signiﬁcant
positive relationship was hypothesized between both education and income and the
dependent measures. The hypothesis regarding education (H10) was partially
supported; in three of the four equations estimating communication differences in the
father-child dyad (Equation 3, 5, and 7) the education of the father was positively
related to father and early adolescent ratings. For the mother-child dyad (Equations
2, 4, 6 and 8) mother’s education was not found to be an explanatory variable. The
hypotheses (H11) that income would be positively related to communication ratings
was not supported: income was found to be unrelated to differences in parent-child

communication.

 

85
The independent variables of gender of child, pubertal age, family income and

parental work status were all found to be insigniﬁcant in explaining differences in

parental ratings of early adolescents on problem-free communication.
WITHIN FAMILY CORRELATIONS

Pearson correlation coefﬁcients were used in order to determine the strength
of the relationship between communication ratings of early adolescents and their
parents within the family (see Table 4-1). As expected (see Chapter 3) early
adolescents’ ratings of their father (openness and problem-free) were not signiﬁcantly
correlated with fathers’ ratings of early adolescents. However, in correlating early
adolescents’ ratings of mothers with mothers’ ratings of early adolescents on
openness, it was found that mothers’ ratings were moderately correlated (r=.349)
with those of early adolescents. The correlation coefﬁcient was positive with a
signiﬁcance level of .01.

Contrary to expectations, early adolescents’ ratings of mother and father on
openness were signiﬁcantly and positively correlated with a coefﬁcient of .591 and
signiﬁcance level of .001. Early adolescents’ ratings of mother and father (problem-
free) were positively correlated (r=.689). Lastly, as expected, mothers’ and fathers’
ratings of early adolescents were positively correlated for openness (r=.287) with a
signiﬁcance level of .01; for problem-free, the correlation coefﬁcient was .491 with

a signiﬁcance level of .001.

CHAPTER 5

CONCLUSIONS AND IMPLICATIONS

CONCLUSIONS

Communication between parents and early adolescents is a complex process
and, therefore, challenging with respect to research. The results of this study
demonstrate the importance of an ecological approach to the study of communication
within families. Factors from the physiological, sociological and psychological
environments of family members and of families as a unit all impact the parent-child
communicative relationship. In applying probabilistic epigenetic theory to the study
of communication, the results of this study contribute to the growing body of research
which addresses issues of context in the study of communicative behavior.

In this chapter, ﬁndings will be summarized within the context of the current
literature, highlighting the importance of an ecological approach in the study of
communication between parents and early adolescents. Although it is difﬁth to

directly compare communication studies, generalizations can be made with respect

86

87

to the implications of the ﬁndings and directions for further research. The discussion

of ﬁndings and implications will be organized according to environmental context.

Physiological Factors

Gender of the early adolescent was not found to be an important factor in this
study. Only in youths’ ratings of mothers on openness was gender a signiﬁcant factor,
with girls reporting more openness on the part of mothers than boys reported. This
ﬁnding is consistent with the literature indicating that girls report better relationships
with their mothers than boys report (Richardson, 1986). Although gender of child
has been found to be important with respect to speciﬁc communication constructs
(i.e., identity exploration, types of information disclosed), it appears that gender of
child is relatively unimportant with respect to the general nature or quality of parent-
child communication.

This study is also consistent with other studies in that differences were found
between mothers and fathers with respect to the parent-child relationship. Early
adolescents rated their mothers higher on openness than they rated their fathers,
reﬂecting the closeness of the mother-child relationship. The closeness of this
relationship was also apparent in the inﬂuence of the child’s perceived acceptance
by mother and father (MACCEPT and DACCEPT). The fact that mothers
signiﬁcantly affected by their child’s perceptions of acceptance, whereas fathers were
not, is consistent with other studies indicating that mothers are more affected by

perturbations in parent-chﬂd relationships than are fathers (Montemayor, 1991).

88

The negative relationship found between pubertal age and early adolescent
ratings of both mother and father is consistent with other studies reporting difﬁculties
in parent-chﬂd relations during the early adolescent period-~particularly at the peak
of puberty. The ﬁnding that ratings of mothers and fathers were not inﬂuenced by
pubertal age of the early adolescent, suggests that the negativism may be one-sided;
that is, adolescents’ views of parents are negatively impacted by pubertal age,
whereas parents’ views of their early adolescents are not significantly impacted by the
child’s extent of physical maturity. Perhaps the view that conﬂict is inevitable during
puberty needs further investigation. It could be that in only a minority of families
the negativism on the part of the early adolescent is transmitted to parents, resulting
in more negative communication patterns.

This study found that the physiological factors had the greatest impact on
early adolescent views and not on parental views. However, only physiological
characteristics of the early adolescent were considered. There were many
physiological factors not examined in this study; variables such as age and general
health of parents, and physical attractiveness of both parents and children should be
included in future studies.

A comment regarding the ﬁndings for early adolescent and parental ratings
of problem-free communication is in order. It is somewhat puzzling, in view of the
fact that early adolescence has been found to be problematic with respect to parent-
child relations, that pubertal age was not found to be a determining factor in either

mothers’ and fathers’ or early adolescents’ ratings on the problem-free scale. This

 

89

ﬁnding may be attributed to: 1) the restricted range of scores in a sample that was
homogeneous and well functioning; and, 2) the reliance on global measures of
conﬂict and problems rather than measures in connection with speciﬁc problematic
instances. Microanalytic analyses of observed interactions over time with more

diverse, heterogeneous samples are needed to address these limitations.

The Sociological Factors
The ﬁnding that family income was not statistically important to differences
in communication ratings is consistent with ﬁndings of studies with families of similar
socioeconomic status. It appears that income is not a determining factor in parent-
child relationships when it is adequate to meet needs. This study does not in any
way negate the ﬁndings for lower socioeconomic groups; that is, that socioeconomic
status is a marker for maladaptive parenting behaviors (Garbarino, 1986).
A contribution of this study is the ﬁnding with respect to parental education.
When examined separately from income, education was found to be statistically
important to the father-child relationship, but not to the mother-child relationship.
It has been suggested (Bronfenbrenner and Crouter, 1982) that indices of social class
should be differentiated in order to determine speciﬁc effects of each parent’s
education and contribution to family income. This study differentiated the effects of
mother’s and father’s educational level and found that father’s, but not mother’s,
education was positively and signiﬁcantly related to differences in communication

ratings. Again, this ﬁnding reﬂects the more intense involvement with their children

90

among mothers as opposed to fathers. It appears that in general, regardless of their
education, mothers communicate more positively and are closer to their children than
are fathers. The greater impact of education for fathers suggests that education
enables fathers to exhibit the ﬂexibility and breadth of perspective important to
communicating with their early adolescent children. The relatively homogeneous
sample is also a factor in this ﬁnding; if this study were repeated with lower
socioeconomic families, education of mothers might also be found to be a statistically
important factor.

The ﬁndings with respect to parental work status suggest that it may be a
marker for position-oriented vs. person-oriented families (Beebe and Masterson,
1986). Position-oriented families rely on the authority and structure of a rigid system
of family roles for maintenance, whereas person-oriented are more ﬂexible in their
role deﬁnitions, adjusting and adapting to the grth and change of their members.
Families in which the mother is a full-time homemaker and the father is the sole
provider are generally more rigid in their role deﬁnitions and behaviors than families
in which both parents are employed. In this study, fathers in families in which
mothers were not employed outside of the home were rated lower in openness by
early adolescents than fathers in families in which mothers were employed; this
ﬁnding supports the theory that child-care is the wife’s domain in these families, and
that fathers may be limited in how they talk to their children. It appears that work
status per se is not a contributing variable to differences in communication ratings,

but rather, the differences are due to value systems and role deﬁnitions within these

.~L~.,1z‘,:;~:-: -

 

91
families. The ﬁnding that mothers who were employed full-time rated children less

open than mothers in other employment categories needs further investigation. This
ﬁnding may point, again, to the less rigid roles in dual-earner families reﬂecting their
more person-oriented approach to relationships as opposed to position-oriented
families; fathers may have a more active parenting role and therefore be closer or
as close as mothers to the children in these families. Another possibility--albeit
unlikely in the light of previous research--is that mothers who are employed full-time
do not have the time necessary to establish as close a relationship as mothers who
spend more time at home. Further study involving communication processes and
behaviors in families of different employment typologies is needed. If parental work
status is a marker for differing values and behaviors, than it is the values, behaviors
and processes that need to be studied in order to facilitate understanding of family

relationships and avoid stereotyping families by employment typologies.

:nic Psychological Factors

An important contribution of this study is the results showing a strong
relationship between paternal job-parent role stress or strain and parent-chﬂd
communication--even in the mother-child dyads, paternal job-parent role stress was
of greater importance than maternal job-parent role stress. This ﬁnding appears to
be inconsistent with the literature suggesting that women are more adversely affected
by balancing job and family roles than men. (Rapoport and Rapoport, 1978). It is

also possible that for this sample of two-parent families, the more traditional view

92

of the importance of father’s employment is reﬂected in the family interaction
patterns. For example, if dad has a bad day at work, it affects the whole family,
whereas mom’s bad day at work becomes secondary to her role as primary nurturer.

In examining the role of family stress as perceived by mother and father on
their respective communication relationships with their early adolescents, it is
apparent that general family stress has a signiﬁcant negative impact on parent-chﬂd
communication. This ﬁnding is consistent with stress theory (McCubbin and
Patterson, 1983) and research. Stress was found to be a greater factor in mothers’
ratings than in fathers’ ratings of children. Perhaps mothers are more affected by
family stress, which may involve relationships within the family, than by work-related
stress. For fathers, work-related stress appears to be a greater inﬂuencing factor in
the way they view their children. If the responsibility for nurturing and maintenance
of relationships within the family falls to mother, than it is logical that family
stressors would have a greater effect on her than on father. It is interesting that
stress was a signiﬁcant factor in the early adolescents’ ratings of their fathers, but not
of their mothers. This may be due also to mother’s role as nurturer; she may strive
to maintain a positive communicative relationship with her child in the midst of
family stress and thus the child is less negatively affected than she is by the stress
that her father is experiencing.

Although the relative contribution of family stress to parent-child
communication was examined in this study, types of stressors were not differentiated.

Thus, questions of the different impact of one type of stressor over another (e.g.

93

sociological versus physiological, normative versus crisis) were not addressed. Future
studies dealing with global stress and parent-chﬂd relations should examine
differential impacts depending on the type of stressor; responses of families over time
need to be analyzed.

An important, but not surprising, ﬁnding was the signiﬁcance of the early
adolescent’s perception of acceptance by mother/father in the parent-child
communication ratings. MACCEPT or DACCEPT was positively and signiﬁcantly
related to communication ratings in six of the eight equations. The importance of
this factor is consistent with the literature linking parental acceptance or warmth to
adolescent self-disclosure, expressiveness and self-esteem. This study also links the
child’s perceived acceptance by the parent to problem-free communication. It
appears that, in addition to the facilitation of openness, problems in communication

are reduced when the early adolescent feels accepted by the parent.

Within Family Diﬂferences

The differences among family members in communication ratings is consistent
with previous research. Early adolescents had a more negative view of parents than
the parents had of their early adolescent children. Mothers were rated signiﬁcantly
higher on openness than fathers, reﬂecting the closeness of the mother-child

relationship. Fathers’ and mothers’ ratings of their child were highly correlated.

94
IMPLICATIONS

Implications for Rmearch

Many studies dealing with communication processes and constructs within
families have, for the most part, ignored issues of context, that is, the environment
in which communication takes place and the impact of that environment on
communication. This study examined a number of environmental inﬂuencing factors
on the parent-child communicative relationship to determine their relative
importance in explaining differences in communication ratings of parents and early
adolescents of each other.

In applying the developmental-contextual model to the study of parent-child
communication, it was necessary to make choices about which variables to include
as predictive or independent variables. This decision was based on previous research
ﬁndings, theory, and, to some extent, the existing data set. In some instances, the
data set resulted in limitations. For example, satisfaction with the work environment
was not included as an explanatory variable because of the limited range of responses
to the one question dealing with satisfaction with work. The ﬁndings of this study
vis-a-vis the work environment support other studies showing that work status is not
as important as attitudes and values about employment. Future research efforts
should concentrate on attitudinal factors.

Repetition of this study or similar ecological studies should be done, either

concentrating on families of lower socioeconomic status or including a larger

95

representation of these families. Although income was not found to be a
contributing variable in this study, the inclusion of families with inadequate incomes
might yield different results. The skewed sample (toward higher education and
income) may have also affected the results for the multiple regression equations
estimating problem-free communication. If lower socioeconomic status is correlated
with problematic parent-child relations, as other studies have indicated, then the
homogeneity of this sample of families may have resulted in the identiﬁcation of only
a few signiﬁcant variables in the attempt to predict the child’s rating of mother and
father on problem-free communication. This relatively highly educated,
homogeneous sample of parents is less likely to exhibit problematic communication
behavior-—such as nagging and insulting their children--than a more diverse
population. The lack of problems may have made it difﬁcult to identify inﬂuencing
factors with respect to problem-free communication.

Research examining communication within families of varying composition is
also recommended. It is important to obtain data from single-parent families and
stepfamilies and compare ﬁndings with respect to family relationships within those
families to two-parent families. Family composition, including numbers of siblings
and spacing of children have also been found to be important in studying family
relationships; the inclusion of these factors in future studies examining parent-child
communication would facilitate understanding of family dynamics.

It is also possible that communication problems may not be adequately

measured using the Olson subscale (problem-free). Some of the items on that scale

96

deal with issues of openness between parents and children, such as the avoidance of
discussing certain topics with a parent or child, as opposed to dysfunctional
communication patterns. Additional research methods, including qualitative analyses
of communication processes in families may be more useful in identifying
problematic communication.

Finally, it is recommended that future studies examine environmental factors
not included in this study. Although the developmental contextual theoretical
approach recognizes the importance of environmental variables in human and family
development and behavior, it is also recognized that it is not feasible to include all
potentially inﬂuential environments in any one study. Although this study included
a community environmental setting that directly affects parents (i.e., parental
employment), community inﬂuences directly affecting the early adolescent (i.e.,

school, relationships with peers) were not a part of this study.

Implications for Education

It should be encouraging to educators and parents alike that there are
commonalities among families with respect to communication between parents and
children during the early adolescent years. For example, if parents realize that
pubertal age has a negative relationship to communicative openness, they will be
better equipped to understand and cope with perturbations in the parent-child
relationships during this time period. Parents also should be encouraged that the '

most important factors affecting parent-chﬂd communication are directly under their

97

inﬂuence (i.e., acceptance of the child by the parent). Parents, especially fathers,
need to be aware of the strong relationship between the child’s perceived acceptance
by parents and positive communication as well as adjustment of the early adolescent.
In addition to awareness, they need to build skills in how to convey their
unconditional acceptance to their child. The fact that acceptance was not
simultaneously determined with the dependent measure of communication, indicates
that perceptions of acceptance by the child are established before the child enters
early adolescence. Thus, it is important that parents demonstrate acceptance and
warmth to their children at early ages.

Educational programs targeted towards early adolescents are also needed.
There is an urgent need for building family life skills in young people. Diminishing
family time, divorce, age-graded societal structure and the increasing dominance of
peers and the media in the lives of children and adolescents has had a detrimental
affect on the quality of parent-child relations. Although some religious institutions
provide educational programs and counseling to help families through the adolescent
years, these programs do not reach large numbers of families and are of varying
quality and effectiveness. Early adolescents can be helped to understand their own
physiological changes, as well as the stresses that they and their parents typically
experience during the early adolescent years. They can be shown how to modify
communicative behavior to maintain positive relations with parents and avoid some

of the turbulence of the adolescent years.

 

98

Implications for Family Policy

The ﬁndings of this study indicate that although mothers are the primary
nurturers of children, fathers in two-parent families play a critical role with respect
to family communication processes. The work environment of the father affects not
only the father-child dyad, but the mother-child dyad as well. It is important that as
a society we support fathers in spending time with their families. Society expects
mothers to put families before employment in terms of time, energies and devotion.
Fathers, on the other hand, frequently have these priorities reversed and society
expects that this is the case. As policy makers strive to build strong families through
legislative and other government efforts, they need to recognize the importance of
allowing and encouraging parents--both fathers and mothersuthe necessary time to
build relationships with their children. Flexible work schedules and time off through
family leave policies should be available to fathers and mothers. As a society we
need to explore all avenues that have the potential to enhance family communication
and relationships--community organizations, religious institutions, parenting classes
and support groups, and time spent with extended families. Working together, we
can help parents and children communicate in ways that facilitate individual grth

and healthy family relationships.

REFERENCES

REFERENCES

Balswick, J .O. and Balkwell, J .W. (1977). Self-disclosure to same-and opposite-sex
parents: An empirical test of insights from role theory. Sociometry, 40(3):282-
286.

Barnes, H.L. and Olson, DH. (1985). Parent-adolescent communication and the
Circumplex model. Child Development, 56, 438-447.

Baumrind, D. (1968). Authoritarian vs. authoritative control. Adolescence, 3, 255-
272.

Beebe, S.A. and Masterson, J. (1986). Family Talk: Interpersonal Communication in
the Family. New York: Random House.

Bird, G.W., and Kemerait, LN. (1990). Stress among early adolescents in two-earner
families. Journal of Early Adolescence, 10(3):344-366.

Bronfenbrenner, U. (1979). The Ecology of Human Development. Cambridge, Mass:
Harvard University Press.

Bronfenbrenner, U. and Crouter, AC. (1982). Work and family through time and
space. In Kamerman, SB, and Hayes, C.D. (Eds.), Families That Work:
Children in a Changing World. Washington, DC: National Academy Press.

Bubolz, M., and Sontag, S. (1990). Human Ecology Theory. Chapter 17 in
Sourcebook of F amily Theories and Methods: A Contextual Approach.

Callan, VJ. and Noller, P. (1986). Perceptions of communicative relationships in
families with adolescents. Journal of Marriage and the Family, 48, 813-820.

99

 

100

Cohen, J ., & Cohen, P. (1975). Applied Multiple Regression/Correlation Analysis for
the Behavioral Sciences. Hillsdale, New Jersey: Lawrence Erlbaum
Associates.

Collins, WA. (1990). Parent-chﬂd relationships in the transition to adolescence:
continuity and change in interaction, affect, and cognition. In Montemayor,
Adams, and Gullotta (Eds), From Childhood to Adolescence. California:
Sage.

Compas, B.E., Howell, D.C., Phares, V. and Williams, RA. (1989). Parent and child
stress and symptoms: An integrative analysis. Developmental Psychology,
25(4):550-559.

Conger, R.D., Conger, K.J., Elder, 6H. and Lorenz, FD. (1992). A family process
model of economic hardship and adjustment of early adolescent boys. Child
Development, 63(3):526-541.

Cozby, RC. (1973). Self-disclosure: A literature review. Psychological Bulletin,
79:73-91.

Crouter, AC, and Crowley, M.S. (1990). School-age children’s time alone with
fathers in single-and dual-earner families: Implications for the father-chﬂd
relationship. Journal of Early Adolescence, 10(3), 296-313.

Dalusio, V.E. (1972). Self-disclosure and perception of that self-disclosure between
parents and their teenage children. Unpublished doctoral dissertation, United
States International University.

Davidson, B., Balswick, J .O., and Halverson, CF. (1980). Factor analysis of self-
disclosure for adolescents. Adolescence, 15(60):947-957.

Dimond, RE, and Munz, DC. (1967). Ordinal position of birth and self-disclosure
in high school students. Psychol. Rep. 21:829-833.

Douvan, E., and Adelson, J. (1966). The Adolescent Experience. New York: Wiley.

 

 

101

Eastman, E., Archer, RP, and Ball, JD. (1990). Psychosocial and life stress
characteristics of Navy families: Family Environment Scale and Life
Experiences Scale ﬁndings. Military-Psychology, 2(2):113-127.

Erickson, EH. (1959). Identity and the Life cycle. New York: International
Universities Press.

Erickson, EH. (1968). Identity: Youth and Crisis. New York: Norton.

Freud, S. (1949). Outline of Psychoanalysis. New York: Norton.

Galambos, N.L, and Maggs, J .L. (1990). Putting mothers’ work-related stress in
perspective: Mothers and adolescents in dual-earner families. Journal of
Early Adolescence, 10(3)313-329.

Garbarino, J ., Schellenbach, C., Sebes, J ., and Associates (1986). Troubled Youth,
Troubled Families. New York: Aldine.

Gekas, V. (1979). The inﬂuence of social class on socialization. Burr, Hill, Nye and
Reiss (Eds.), Contemporary Theories About the Family. New York: Macmillan
Publishers.

Gilbert, S. (1976). Self-disclosure, intimacy and communication in families. The
Family Coordinator, 25:221-31.

Glenwick, D.S. and Mowrey, J .D. (1986). When parent becomes peer: Loss of
intergenerational boundaries in single parent families. Family Relations, 35,
57-62.

Gold, D., and Andres, D. (1978). Developmental comparisons between ten-year-old
children with employed and nonemployed mothers. Child Development, 49,
75-84.

 

102

Gottlieb, G. (1970). Conceptions of prenatal behavior. In R. Aronson, E. Tobach,
D.S. Lehrman, and J .S. Rosenblatt (Eds.), Development and Evolution of
Behavior: Essays in Memory of T. C. Schneirla. San Francisco: Freeman.

Grotevant, H.D. and Cooper, C. (1985). Patterns of interaction in family
relationships and the development of identity exploration in adolescence.
Child Development, 56, 415-428.

Hauser, S.T., Powers, S.I., Noam, G.G., and Jacobson, AM. (1984). Familial
contexts of adolescent ego development. Child Development, 55, 195-213.

Hausman, J .A. (1978). Speciﬁcation tests in econometrics. Econometrica, 46:1251-
1271.

Hill, J .P. (1980). Understanding Early Adolescent: A Framework. Chapel Hill:
Center for Early Adolescence, University of North Carolina.

Hill, J .P. and Holmbeck. G. (1987). Disagreements about rules in families with
seventh graders. Journal of Youth and Adolescence, 5, 31-44.

Hill, J ., Holmbeck, G., Marlow, L., Green, T. and Lynch, M. (1985). Menarcheal
status and parent-child relations in families of seventh grade girls. Journal of
Youth and Adolescence, 14, 314-330.

Hoffman, L.W. (1977). Changes in family roles, socialization and sex
differences. American Psychologist, 32, 8, 644-657.

Hoffman, L.W. (1989). Effects of maternal employment in the two-parent Family.
American Psychologist, 44(2), 283-292.

Hunter R., and Youniss, J. (1982). Changes in functions of three relations during
adolescence. Developmental Psychology, 18, 806-811.

 

103

Jacob, T. (1974). Patterns of family conﬂict and dominance as a function of child
age and social class. Developmental Psychology, 1021- 12.

J ourard, SM. (1971). Self-disclosure: an Experimental Analysis of the Transparent Self.
New York: Wiley.

Johnson, C.L., and Johnson, RA. (1977). Parenthood, marriage, and careers:
Situational constraints and role strain. In R. Rapoport and RN. Rapoport
(1980), Dual-career Couples. Beverly Hills, California: Sage.

Keith, J ., Nelson, C., Potter, P. Thompson, C., and Schlabach, J. (1992). Michigan
Early Adolescent Survey 11, Research Report 504. Michigan State University
Agricultural Experiment Station, East Lansing, MI.

Kohn, ML. (1969). Class and Conformity: A Study in Values. Homewoood, IL.:
Dorsey.

Kohn, ML. (1979). The effects of social class on parental values and practices. In
Reiss, D., and Hoffman, HA, The American Family: Dying or Developing.
New York: Plenum.

Komarovsky, M. (1974). Patterns of self-disclosure of male undergraduates. Journal
of Marriage and the Family, 36(4):677-686.

Kraft, L.W., and Vraa, CW. (1975). Sex composition of groups and pattern of self-
disclosure by high school females. Psychol. Rep. 37:733-734.

Lerner, R.M., Lerner, J .V., and Tubman, J. (1990). Organismic and contextual bases
of development in adolescence: A developmental contextual view. Adams,
G.R., Montemayor, R., and Gullotta, T.P. (Eds.), Biology of Adolescent
Behavior and Development. California: Sage Publications.

Lewis-Beck, M.S. (1980). Applied Regression: An Introduction. Beverly Hills: Sage
Publications.

 

104

McCubbin, HI. and Patterson, J.M. (1983). The family stress process: The
double ABCX model of adjustment and adaptation. Marriage and Family
Review, 6(1,2): 7-37.

Montemayor, R. (1983). Parents and adolescents in conﬂict: All families some of
the time and some families most of the time. Journal of Eariy Adolescence, 3,
83-103.

Montemayor, R. (1984). Maternal employment and adolescents’ relations with
parents, siblings, and peers. Journal of Youth and Adolescence, 13(6):543-557.

Montemayor, R. Adams, GR, and Gullotta, T.P. (1990). From Childhood to
Adolescence. California: Sage Publications.

Mortimer, J .T., and London, J. (1984). The varying linkages of work and family. Pp.
20-35 in P. Voydanoff (Ed.), Work and Family: Changing Roles of Men and
Women. Palo Alto, CA: Mayﬁeld.

Nelson, CS. (1985). Toward an ecology of early adolescent sex role development:
Attitudes and behaviors. Doctoral dissertation, Michigan State University,

East Lansing, MI.

N iemi, PM. (1985). Family interaction patterns and the development of social
conceptions in the adolescent. Journal of Youth and Adolescence, 17(5):429-
443.

Noller, P., and Bagi, S. (1985). Parent-adolescent communication. Journal of
Adolescence, 8:125-144.

Noller, P., and Callan, V.J. (1986). Adolescent and parent perceptions of family
cohesion and adaptability. Journal of Adolescence, 9:97-106.

Norrell, J .E. (1984). Self-disclosure: Implications for the study of parent-adolescent
interaction. Journal of Youth and Adolescence, 13(2):163-177.

105
Offer, D. (1969). The Psychological World of the Teenager. New York: Basic Books.

Olson, D.H., Portner, J ., and Bell, R. (1982). FACES II. University of Minnesota,
St. Paul, MN.

Olson, D.H., Russell, CS. and Sprenkle, DH. (1983). Circumplex model of marital
and family systems: VI theoretical update. Family Process, 22, 69-83.

Olson, D.H., Sprenkle, DH, and Russell, CS. (1979). Circumplex model of marital
and family systems: I cohesion and adaptability dimensions, family types and
clinical applications. Family Process, 18, 3-15.

Orthner, D.K. (1990). Parental work and early adolescence: Issues for research and
practice. Journal of Early Adolescence, lO(3):246-260.

Papini, D., and Sebby, R. (1987). Adolescent pubertal status and affective family
relationships: A multivariate assessment. Journal of Youth and Adolescence,
16, 1-15.

Paulson, S.E., Koman, J .J ., and Hill, J .P. (1990). Maternal employment and parent-
child relations in families of seventh grade girls. Journal of Early Adolescence,
10(3):279-296.

Piaget, J. (1970). Piaget’s Theory. In P.H. Mussen (Eds.), Carmichael’s
Manual of Child Psychology, 3rd ed. New York:Sage.

Piotrkowski, CS. (1979). Work and the Family System. New York: MacMillan.

Piotrkowski, C.S., Rapoport, R.N., and Rapoport, R. (1987). Families and work. In
M.B. Sussman and SK. Steinmetz (Eds.), Handbook on Maniage and the
Family (pp. 251-283). New York: Plenum.

 

106

Rapoport, R., and Rapoport, RN. (1978). Three Generations of Dual-Career
Family Research. In R. Rapoport and RN. Rapoport (Eds.), Dual Career
Couples (1980). Beverly Hills, Calif: Sage.

Richardson, R.A., Abramowitz, R.H., Asp, EB, and Petersen, A.C. (Nov 1986).
Parent-child relationships in early adolescence: Effects of family structure.
Journal of Marriage and the Family, 48:805-811.

Rivenbark, W.H. (1971). Self-disclosure patterns among adolescents. Psycho. Rep.
28:35-42.

Savin-Williams, RC, and Small, S.A. (1986). The timing of puberty and its
relationship to adolescent and parent perceptions of family interactions.
Developmental Psychology, 22(3):342-347.

Scarr, 8., Phillips, D., and McCartney, K. (1990). Working mothers and their
families. In Chess, S. and Hertzig, M. E., Eds. Annual Progress in Child
Psychiatry and Child Development. New York: Brunner/Mazel, 261-278.

Schlabach, J .H. (1989). Income, work, and spending behavior of early adolescents.
Master’s thesis, Michigan State University, East Lansing.

Silverberg, SB, and Steinberg, L. (1987). Adolescent autonomy, parent-adolescent
conﬂict, and parental well-being. Journal of Youth and Adolescence, 16(3):293-
311.

Steihberg, L.D. (1981). Transformation in family relations at puberty.
Developmental Psychology, 17, 933-840.

Steinberg, L.D., and Hill, J .P. (1978). Patterns of family interaction as a function of
age, the onset of puberty, and formal thinking. Developmental Psychology,
14(6):683-684.

Steinberg, L. (1987). Impact of Puberty on family relations: Effect of pubertal status
and pubertal timing. Developmental Psychology, 23, 451-460.

 

 

107

Thomson, B., and Vaus, A. (1986). The importation, transmission, and moderation
of stress in the family system. American Journal of Community Psychology,
14(1):39-57.

Wiebe, B., and Williams, J .D. (1972). Self-disclosure to parents by high school
seniors. Psycho. Rep. 31:690.

Yang, M., and Hwang, K. (1980). The vedge model of self-disclosure and its
correlates. Acta Psychol. Taiwan, 22:51

Youniss, J ., and Smollar, J. (1985). Adolescent Relations with Mothers, Fathers, and
Friends. Chicago: University of Chicago Press.

Zellner, A. (1963). Estimators of seemingly unrelated regressions: Some exact ﬁnite
results. Journal of the American Statistical Association, 58:977-992.

APPENDIX

 

108

 

.88 88. 2.8050 33. 8a .2552. .6 258a 3120 a. 51 2: .3: 96:3 3 .838 a 895 :2? Samoan. 8- 2:. 88. .568 8.. 8a 823.: 68.698 05. .8

... ....8. 0.58.... .1... ...... . ...“.qu .351» -.f a». E .b ﬁzmoéaosgho .858“.
«S u ooaﬁoam co Sum ....m :85 r. End. .80 u R
82. v a 8... v a so... v o
an: 8a.: Gog
$9.... - :25 + 88.4
5. n oooaoam no sum .3 ”8... u and as... u am
8.. v a :5... v a
an; 325
$2... . 88m
:3 u 239m Co Sam on 385 u and “as u R
n8... v Q So... v a ~86 v a
5.3 a; 923
was... . 38° + 83a
98 u 2.86am no arm on 38... u an...“ 59 u so.
n8... v a #8.: v a
God 533
33... - Ewan
So u soaoam co Sam on .2... u Em...— "35 u am
on... v a so... v a
Gog A86
33... + 53
HS ... oaaoom co Sam on "3.: u and "3.: u a...
8... v a #8... v a
Gas no.8
$85 - 88m
28 u 333m no saw no :85 a sad Home u «a
W85 v a 33 v a so... v a
83 63 98V
285 - «an... + «3%
ood u 359m co Sam on 385 u sans “one u am
83. v a n8... v a 83. v a ~85 v a
63 $3 $5 63v
885 .

 

38:8 8a «80. Jcoua E a, ..E:

 

E1562 mam

$5an a em

8550: o em

05.58 m em

.5020 4 am

3.30 m am

02020 N ...m

 

$5.9- .vmvmd + Soﬁa

 

aul—3.3— -.Q—Dmv canny—way— Baa—ED waEoom mace: Eozmzram EzEuEaEEeU u-EQEEQ .~.< o—Auﬁ.

109

.98" 88.. 380.50 “ma. «0: .oéamo: .5 0350a .35? 3 a»: on. 35 05:3 0. :83.— m_ 82: :25 Sam—mega 2a 05 83“ =S-o:0 5.. PS mos—aim 02.592 2F .8
"o .20 B 25 w. = no Boo 5% . 50:0.— 2 22:53 2a 3355qu E Eons:

 
      

   

     

and v a 86 v a
20.8 and
«.de + wag - n 05.502 w em
as v .2 85 v .2 3o v .2 8o v a 86 v .2
3.8 85v 305 Ga: 82V
885 + 808 - 885 + 286 - 38d - u 03.500 5 em
85 v a 2.9 v a on... v a 085 v a
83 600 808 $00
$25 - 3.8 - 98.0 + $.25 - u 005.502 o am
82 v a as v .2 Rod v a
and 908 900
$88 + $85 - 83¢ - N 02500 m cm
n .2020 10m
cod v a
$9
:86 - n .2930 m em
85 v a as v .2 Rod v a
GEO 32.8 3.8
885 - owed . «88 - n 02020 N am
80o v a and v a
$5 5.8
.o - 0o - n 0950 2 am

 

 

Zn mm MNQZm—O ODQNQQ 8MP; mm—MHEQ
ﬂj

 

4:00 T< 030,—.

HICHIGRN STQTE UNIV. LIBRQR

ES

    

| I IN IHHIIH Ill

31293008853297