 

, 1...“,

5 v:

a u
Sht.

6.1;. 1:

Au 0'7

 

. .nv...i‘ou Jopln

3
53

1g

1.

“mania. n

:- lip»

 

‘ be”;

mm

,2?
L ' ()(\L’
59W/995?

This is to certify that the
dissertation entitled

RELATIONSHIPS BETWEEN PARENTAL ACCEPTANCE-
REJECTION, FAMILY FUNCTIONING AND
DISORDERED EATING IN COLLEGE-AGED FEMALES

presented by

TIANNA HOPPE-ROONEY

has been accepted towards fulﬁllment
of the requirements for the

PhD. degree in Family and Child EcologL

 

Mi Maw, p141

Major Professors Signature

Aggst 25, 2004

 

Date

MSU is on Allinnadve Action/Equal Opportunity Institution

- 4—-—-—-- —o—-o-—-p..---.-o—-.—----n-o-u-.-o-c-u---a-o-n-o-u—o. A-—‘-.-o-o-o-.-.-o-o----—

 

LIBR HY

Michigan State
University

———

 

 

 

 

PLACE IN RETURN Box to remove this checkout from your record.
To AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE DATE DUE DATE DUE

 

 

m n (2 {mg
IP31 H W

 

 

 

 

 

 

 

 

 

 

 

 

 

6’01 cJCIRC/DateDue.p65-p.15

 

O

RELATIONSHIPS BETWEEN PARENTAL ACCEPTANCE-REJECT ION,
FAMILY FUNCTIONWG AND DISORDERED EATING IN
COLLEGE-AGED FEMALES

By

Tianna Hoppe-Rooney

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

2004

ABSTRACT

RELATIONSHIPS BETWEEN PARENTAL ACCEPTANCE-REJECTION,
FAMILY FUNCTIONING, AND DISORDERED EATING IN
COLLEGE-AGED FEMALES

By

Tianna Hoppe-Rooney

The purpose of this study was to better understand the relationships between
parental acceptance-rejection, family functioning, and disordered eating. Specifically the
study set out to address the research question: Are parental acceptance-rejection and
family functioning each related to disordered eating symptomatology? Variables
examined in the study include maternal and paternal acceptance—rejection, and family
functioning as independent variables and eating attitudes and behaviors as the dependent
variable. Purposive sampling on a large university campus achieved a sample of 834
female participants between the ages of 18 and 25 years, enrolled as undergraduates.

In comparing the disordered eating and non—disordered eating groups in the study
as differentiated by the EAT-26, it was found that statistically signiﬁcant differences
exist between the two groups on the measures of parental acceptance—rejection and family
functioning. The disordered eating group reported greater levels of maternal rejection,
paternal rejection, and unhealthy family functioning as compared to the non-disordered
eating group.

The variables with the greatest impact on the dependent variable of disordered
eating included maternal acceptance-rejection, family problem solving, and family

behavior control.

Copyright by
TIANNA HOPPE-ROONEY
2004

ACKNOWLEDGEMENTS

Although I have never considered myself a runner, an analogy of a race comes to
mind when reflecting on my graduate work. Many people in my life have helped to train
me for the sprint as well as keep me going when the race seemed too difﬁcult to ﬁnish.
My crossing the ﬁnish line could not have been achieved without the constant attention,
support, and encouragement from my husband and biggest supporter, Todd. You picked
me up when my mind was tired and nourished me in ways that meant so much to me
along the intense journey. You believed, more than I at times, that I could succeed at
running this race. Thank you.

A tremendous amount of gratitude is delivered to my committee chair, mentor
throughout my entire graduate training, and coach in many ways, Dr. Marsha Carolan.
Your encouragement and persistency in reminding me of my abilities always seemed to
pick me up when I was feeling exhausted or unsure of myself. Thank you.

My doctoral committee members, Dr. Thomas Luster, Dr. Kelly Klump, and Dr.
Esther Onaga, contributed their expertise in ways that sincerely fostered both my learning
and my conﬁdence as a researcher and scholar. My selection of these ﬁne professors to
assist me in my journey was truly what made this process one of skill building and
collaboration. Thank you all.

When beginning to tire or become frustrated along the way, I could always count
on Whitney Brosi and Matt Brosi to remind me that everyone needs a break to refresh
their mind and body. Our weekend adventures indirectly contributed so much to my
growth process as a developing professional. I appreciated learning the importance of

balance from you both. Thank you.

iv

TABLE OF CONTENTS

LIST OF TABLES ........................................................................... vii-viii
LIST OF FIGURES .......................................................................... ix
CHAPTER 1
INTRODUCTION
Statement of the Problem ........................................................... 1
Purpose of the Study ................................................................. 2
Theoretical Frameworks ............................................................ 3
Human Ecological Theory ................................................ 4
Symbolic Interactionism ................................................... 10
Theoretical Map ...................................................................... 15
Research Questions .................................................................. 16
Hypotheses ............................................................................ l7
Conceptual Model .................................................................... 18
Methodology .......................................................................... 19
Overview of the Dissertation ....................................................... 19
CHAPTER 2
REVIEW OF LITERATURE
Introduction ........................................................................... 20
Disordered Eating and Familial Environment .................................... 20
Early Family Typologies ................................................... 20
Content versus Process Driven Research ................................. 21
Perceived Family Functioning and Environment in Eating
Disordered Families ..................................................... 24
Parental Inﬂuences .......................................................... 29
Construct of Rejection ...................................................... 33
Parental Acceptance-Rejection Framework .............................. 36
Supporting PARTheory .................................................... 40
CHAPTER 3
METHODOLOGY
Introduction ........................................................................... 45
Research Questions and Hypotheses .............................................. 46
Conceptual Model ................................................................... 48
Conceptual and Operational Deﬁnitions .......................................... 48
Sampling Procedure .................................................................. 50
Variables .............................................................................. 52
Instrumentation ....................................................................... 54
Data Collection ....................................................................... 62
Data Analysis ......................................................................... 64

TABLE OF CONTENTS (cont’d)

CHAPTER 4
RESULTS
Introduction ........................................................................... 66
Data Collection and Sample Recruitment ........................................ 66
Demographic Sample Characteristics ............................................. 68
Variable Distributions across the Sample ......................................... 70
Hypotheses ............................................................................ 79
Hypothesis 1 ................................................................. 8O
Hypothesis 2 ................................................................. 83
Hypotheses 3 and 4 ......................................................... 84
Hypothesis 5 ................................................................. 100
CHAPTER 5
DISCUSSION
Introduction ........................................................................... 102
Overview of Findings ............................................................... 103
Conceptual Models .................................................................. 104
Original Conceptual Model ................................................ 104
Revised Conceptual Model ................................................ 105
Evaluation of Research Questions ................................................. 106
Research Question 1 ........................................................ 106
Research Question 2 ........................................................ 107
Research Question 3 ........................................................ 108
Research Question 4 ........................................................ 110
Research Question 5 ........................................................ 113
Limitations ............................................................................ l 16
Sample ........................................................................ 1 16
Self Report Measures ....................................................... 117
Sequencing of the Questionnaires ......................................... 118
Undifferentiated Eating Disorder Symptomatology .................... 119
Clinical Implications ................................................................. 120
Treatment on the College Campus ........................................ 120
Treatment within the Family ............................................... 123
Future Directions ..................................................................... 126
APPENDICES
APPENDD( A: Informed Consent ................................................. 130
APPENDIX B: Demographic Questionnaire ..................................... 131
APPENDIX C: Eating Attitudes Test (EAT-26) ................................. 133
APPENDD( D: Parental Acceptance-Rejection Questionnaire- Mother. 134
APPENDIX E: Parental Acceptance-Rejection Questionnaire- Father ....... 139
APPENDIX F: Family Assessment Device (FAD) .............................. 144
REFERENCES ................................................................................ 149

vi

LIST OF TABLES

Table 1: Abbreviated Model - Variables and Corresponding Instruments ............ 61
Table 2: Expanded Model - Variables and corresponding Instruments ................ 62

Table 3: Research Question with Corresponding Hypothesis and Data Analysis. 64

Table 4: Summary Demographic Characteristics of Sample ............................ 70
Table 5: Frequency Data for the FAD General Functioning Scale ..................... 73
Table 6: Maternal Parental Acceptance-Rejection Data: Raw and Z-scores .......... 75
Table 7: Paternal Parental Acceptance-Rejection Data: Raw and Z-scores ........... 77
Table 8: Group Statistics for PAR Variables included in T-tests ....................... 82
Table 9: Homogeneity of Variance and T-test Analyses for PAR Variables .......... 82

Table 10: Group Statistics for Family Functioning Variable included in T-test. . 83
Table 11: Homogeneity of Variance and T-test Analyses for Family Functioning...83
Table 12: Correlation Data for the 3 Primary Independent Variables .................. 85

Table 13: Summary of Regression Analysis using the Enter Method for the Study’s
Abbreviated Model Predicting Eating Attitudes and Behaviors... ... .. .....90

Table 14: Summary of ANOVA for Study’s Abbreviated Model using
Enter Method ...................................................................... 90

Table 15: Summary of Regression Analysis using the Step-wise Method for the
Study’s Abbreviated Model Predicting Eating Attitudes and Behaviors. . .92

Table 16: Summary of ANOVA for Abbreviated Model using Step-wise Method. 92
Table 17: Correlation Matrix for Study’s Independent Variables ...................... 95

Table 18: Summary of Regression Analysis using the Enter Method for the
Study’s Expanded Model Predicting Eating Attitudes and Behaviors. 97

Table 19: Summary of ANOVA for Study’s Expanded Model using
Enter Method ...................................................................... 98

vii

LIST OF TABLES (cont’d)

Table 20: Summary of Regression Analysis for the Study’s Follow-up Model ...... 98

Table 21: Summary of ANOVA for Study’s Follow-up Model using
Enter Method ...................................................................... 98

Table 22: Frequencies for Speciﬁc EAT-26 Questions .................................. 122

viii

LIST OF FIGURES

Figure 1: Theoretical Map ................................................................... 16
Figure 2: Conceptual Model ............................................................ 18, 48, 104
Figure 3: Distribution of EAT-26 Composite Scores .................................... 71
Figure 4: Distribution of FAD General Functioning Scores ............................. 74
Figure 5: Histogram for Raw Composite of MPARQ .................................... 76
Figure 6: Histogram for Z-score Composite of MPARQ ................................ 76
Figure 7: Histogram for Raw Composite of FPARQ .................................... 78
Figure 8: Histogram for Z-score Composite of FPARQ ................................. 79
Figure 9: Abbreviated and Expanded Regression Models ............................... 86

Figure 10: Histogram of Residuals using EATCOMP as Dependent Variable ....... 87

Figure 11: Normal P-P Plot of Residuals using EATCOMP as
Dependent Variable .............................................................. 88

Figure 12: Histogram of Residuals using EATLOG as Dependent Variable ......... 88
Figure 13: Normal P-P Plot of Residuals using EATLOG as Dependent Variable. 89

Figure 14: Revised Conceptual Model ..................................................... 105

ix

CHAPTER 1
Introduction
In this chapter, the need for research linking the perception of parental
acceptance-rejection and family functioning with eating disorders will be outlined. Two
theoretical models, Human Ecology Theory and Symbolic Interactionism, will be
discussed in detail, providing a foundation for the studies methodology. A theoretical
map will be introduced to provide clarity in conceptualizing the two theoretical
frameworks being incorporated into the project. Next research questions for the project
are posed, followed by their corresponding hypotheses. A conceptual map of the
relationships between variables is presented preceding a brief discussion of the project’s
methodology, which will be discussed more in-depth in chapter 3. Chapter 1 concludes
with an overview of the content within the project’s remaining chapters.
Statement of the Problem
Post-puberty estimates of eating disorder rates in the United States indicate that 5-
10 million females and 1 million males struggle with diagnosable or sub-clinical
conditions (Crowther, Wolf, & Sherwood, 1992; Fairbum, Hay, & Welch, 1993; Shisslak,
Crago, & Estes, 1995). Statistics focusing more speciﬁcally on rates of clinical diagnosable
eating disorders ﬁnd 2 million females (3%) meeting these criteria in the United States
(Mussell, Binford & Fulkerson, 2000) with incidence rates for anorexia nervosa ranging
between .5%-1.0% and l.0%-3.0% for bulimia nervosa (American Psychiatric
Association, 1994). In the college age population the rates reach as high as 4% to 13%

(Halmi, Frank & Schwartz, 1981; Smith & Thelen, 1984). It is estimated that

approximately 90% of individuals seeking mental health services for eating disorders are
female (Fairburn & Beglin, 1990).

Further, peak onset for anorexia nervosa is 17 years of age (American Psychiatric
Association, 1994), implicating the family system in the etiology or maintenance of the
illness as the child is assumed to still be living in the home environment. Although the
incorporation of disordered eating into the family sciences literature has increased
considerably over the past two decades, a study looking at the relationship between the
presence of disordered eating and perception of both parental rejection and unhealthy
family functioning is absent. The proposed study calls for a remedy to this problem of
insufﬁcient knowledge regarding these relationships.

Purpose of the Study

The current study aimed to explore the relationship between eating attitudes and
behaviors with perception of parental acceptance-rejection and family functioning.
Greater understanding of these possible linkages serves to outline future therapeutic
interventions to be used with individuals seeking treatment for eating disorder behaviors.
It is well known that eating disorders are multi-faceted and multi-dimensional
phenomena, whereby numerous variables account for the presence and maintenance of
symptomatology. A person-context model was used in this study’s research design
whereby characteristics of both the person and the environment were examined as to their
effects on the individual’s development, speciﬁcally their eating attitudes and behaviors.
The accumulation of new knowledge in this area, including the incorporation of both
parental rejection and family functioning, allows for increased understanding of the

multitude of variables lending to an individual’s eating attitudes and behaviors.

Theoretical Frameworks

Two primary theoretical frameworks have been identiﬁed to guide the
conceptualization of the proposed research. Human ecology theory and symbolic
interactionism connect individual functioning to the immediate familial environment.

Human ecology theory anchors the proposed study through the incorporation of
contextual family factors into the key variables: parental rejection and family functioning.
Human ecology theory, as developed by Bronfenbrenner (1979), contends that a person’s
ecology includes four interacting systems including the microsystem, mesosystem,
exosystem, and macrosystem. The proposed study focuses on the microsystem of the
family environment, examining the variables of rejection and functioning. The family
microsystem is rich with communication and reciprocal interactions between members.
Study participants’ perceptions of parental rejection/acceptance and family functioning
will be used to understand their family context and identify if these two constructs
inﬂuence the presence of disordered eating symptomatology. Bronfenbrenner’s concept
of proximal processes as embedded within the microsystem creates a context for meaning
making and perceptual interpretation, inviting symbolic interactionism as a second
guiding theory into the study.

According to symbolic interactionism, perceptions are extracted through
reciprocal interactions occurring within the microsystem. These deﬁnitions, as interpreted
by the individual family members, connect human ecology theory to symbolic
interactionism. The notion of deﬁning the situation, according to symbolic interactionism
theory, occurs at the microsystemic level including the family context. Each individual

family member interprets or makes sense of a situation or event. It is critical however, to

view this individual processing as part of the mutual interactive sequences that occur in
the context of families. The perception of parents’ level of acceptance or rejection as well
as the perception of the family’s functioning will be assessed through the experiences of
the study participant. Following is a discussion of each of the study’s guiding theories,
human ecology theory and symbolic interactionism.

Human Ecological Theory

Urie Bronfenbrenner (1979) described the basis of an ecological approach to
human development as including “the interaction between individual and environment.”
An orientation to context and the interaction between an organism and its environment
deﬁnes an ecological perspective as looking beyond the individual organism to its
environment for questions and explanations about the organism’s behavior, functioning,
and development (Bronfenbrenner, 1979; Griffore & Phenice, 2003).

The incorporation of contexts with the developing person makes up the
foundation of human ecology theory. Human ecology theory draws from biological,
behavioral, and social sciences with a focus on the conditions and processes, which
contribute to the shaping of human development (Bronfenbrenner, 1995, Bubolz &
Sontag, 1993; Bubolz, Eicher, & Sontag, 1978). Classic human ecology theory
conceptualizes four environments: the microsystem, mesosystem, exosystem, and
macrosystem, arranged in a concentric circle formation (Bronfenbrenner, 1979). The
proposed study looks speciﬁcally at the microsystem.

The microsystem is deﬁned as the actual setting in which the individual
experiences and creates day-to—day reality including the places inhabited, the people

living with them or surrounding them, and the activities they participate in together

(Bronfenbrenner, 1979; Bubolz & Sontag, 1993). The current investigation uses the
perceptions derived from the interactions between the study participant and her family
over time. The microsystems in which study participants are assumed to be involved
include family-of-origin and the university setting; however, only family-of-origin
perceptions are being actively investigated. The university and peer microsystems are not
included in the present study.

Human ecology theory compliments family systems theory through their
conceptualizations of how and why an individual functions (or doesn’t function). For
instance, Minuchin, Rosman, and Baker (p. 21, 1978) state, “the psychological unit is not
the individual; it is the individual in her signiﬁcant social contexts.” The “signiﬁcant
social context” being studied here is the perception of the familial environment.

Human ecology theory uses a process orientation rather than a content driven
analysis, whereby all systems and environments are assumed to interact. Four major
components included in this process perspective include perceiving, valuing, decision-
making, and spacing (Freedman Melson, 1980). For the purposes of this study, primary
attention will be paid to the concept of perceiving.

Perceiving is a process whereby environmental information is registered,
organized and made available for use by the system (Freedman Melson, 1980). It is a
process of deriving meaning from environmental stimuli. For instance, a child uses cues
from his or her parents, ﬁltering them through their information processing avenues.
Perception involves interpretation based on previous experiences and future expectations.
Each family member relates to the family as it appears through the prisms of his or her

own perceiving process (Freedman Melson, 1980). Using human ecology theory, the

incorporation of context into understanding individuals’ processing enhances researchers’
abilities to study how and why people think, behave, and feel the way they do.

Since the original formulation of human ecology theory in the 19708,
Bronfenbrenner has more recently gone on to shift his language to include a
“bioecological paradigm” (Bronfenbrenner, 1995). Bronfenbrenner describes the
bioecological concepts as “latent” within the original human ecology theory, but more
recently he strives to make them explicit in his literary works. The bioecological
paradigm uses a processperson-context-time (PPCT) model. All four of the dimensions
included in the PPCT paradigm are considered to be interconnected and interdependent.
A brief description of each follows.

_C_3_o_n_t_e_x_t is speciﬁed to include both an individual’s biopsychological
characteristics and the environments in which she lives and grows. This multifaceted
inﬂuence approach on development views the person as context, including her genetic,
biological, and psychological processes. The tiﬂ: dimension of the PPCT model includes
historical and cultural events as well as life cycle stages that impact an individual’s
development. Finally, the M dimension refers to interactions that take place over
time between an individual and signiﬁcant others, objects or symbols in their life.

Embedded within the bioecological model are proximal processes, which are used
as templates for studying human development and interaction. Proximal processes consist
of reciprocal interactions over time between a human organism and the various
components of his or her external environment, such as persons, objects, and symbols
(Bronfenbrenner, 1999). An example of a proximal process includes the interaction

between a baby and mother as the mother feeds and strokes the ﬁne hair of the infant

while the baby reciprocates the attention by falling to sleep, sucking, smiling, or touching
the mother’s skin. A key piece to understanding proximal processes is to acknowledge
the interaction takes place over time. Interactions must continue long enough to become
increasingly complex, such as those between parent and child (Fenichel, 2002).

Proximal processes form a foundation whereby a reciprocal relationship is
fostered between dyads in a family system. In accordance with family systems’ theory,
the family consists of interdependent individuals whose perceptions and resulting
behaviors mutually affect one another. A meta-analysis by Rothbaum and Weisz (1994)
found children’s contribution to the relationship between parental rejection and behavior
problems is at least as strong as the parents’, which implies a bi-directional, reciprocal
relationship. Perceived rejection appears to elicit a variety of aversive personality and
behavioral characteristics that, in turn, provoke the susceptible parent to further reject the
child. The child responds to the further rejection, and the exchange eventually leads to
development of a mechanism to cope with the pain (Campo & Rohner, 1992).

This is a process view of the psychological concept of perception and family
system’s concept of interdependence. Bronfenbrenner and Ceci (1994) state that these
processes involve content and, at the beginning, this content is in external to the person.
Over time, development consists of involving the interaction between the organism and
her environment, whereby “the external becomes internal and becomes transformed in the
process” (Bronfenbrenner & Ceci, 1994, p. 575). This is the aspect of the process, which
the current study focuses on, the internalization of the perceived external. In other words,
a daughter’s incorporation of how she perceives her parents to accept or reject her into

the many facets of her life including her attitudes and behaviors involving food. The

extension of this process not included in the current study, but important to note is the
idea that the organism than begins to change her environment over time, whereby “the
internal becomes external” (Bronfenbrenner & Ceci, 1994, p. 575). This reciprocal
interaction typiﬁes proximal processes within the bioecological model.

Bronfenbrenner postulates in the bioecological paradigm that proximal processes
may serve to buffer against environmental differences in developmental outcomes
(Bronfenbrenner, 1995). In other words, if a child were to interpret the interaction
between herself and her primary caregiver as satisfying, emotionally fulﬁlling, and
communicative of acceptance, it is predicted that this relationship will inﬂuence the
presence of disordered eating attitudes and behaviors on the part of the child.
Bronfenbrenner goes on to state,

“What is most revealing about proximal processes, however, is not the gain in

predictive power that they provide, but their substantive and theoretical

signiﬁcance as the mechanisms of organism-environment behavioral interaction
that drive development, and the profound ways in which these mechanisms are
affected by characteristics of the developing person and of the environmental

context in which the interaction takes place (Bronfenbrenner 1995, p. 626).”

Using the PPCT model as part of the larger bioecological and human ecology
frameworks, this project theorizes the inclusion of the following variables to fulﬁll each
of the model dimensions. First, process will be measured indirectly through the
assessment of both parental acceptance-rejection and family functioning. It is assumed
that the proximal processes at the base of an individual’s relationship with her parents
will factor into her interpretation and classiﬁcation of their relationships. Bronfenbrenner
(1995) has called for research looking at different types of dyads on the nature and power

of the proximal processes that take place, and the resultant kinds of developmental

outcomes fostered. The current study looks at the mother-child and father-child dyads

with the quality of proximal processes implied in the study’s independent variables of
parental acceptance-rejection and family functioning.

The sample will consist of females between the ages of 18 and 25, indicating the
“person” dimension of the study. Additionally, the score derived from the eating attitudes
test will indicate a component of this dimension as it is theorized to include psychological
processes. Bronfenbrenner dichotomizes two types of person characteristics including
biopsycholgoical resources and directional dispositions. This study draws upon the latter,
which highlights the person as an active agent who selects stimuli to be responsive to.
The former refers to an individuals abilities, temperament, and personality which are not
included as variables in this study.

The context of the study’s participants is being measured through several
variables including family income (socio-economic status), birth order, parent’s
relationship status, and family functioning assessment. Further, each participant is
assumed to be immersed in the university environment, making up a physically and
culturally different context compared to that of their family’s home environment.

Finally, the time dimension of the model is linked to the context of the
participant’s lives. Based on their age and enrollment in college courses it is assumed that
the majority of the study participants will be in the launching stage of their life cycle
according to Duvall, connoting their developmental phase of broaching adulthood
(Strong, DeVault, Sayad, & Cohen, 2001).

According to the bioecological paradigm and human ecology theory, the inclusion
of both beliefs and behaviors in a research design will enhance the explanatory power of

analytic models (Bronfenbrenner, 1995). In the project at hand, perceptions of parental

acceptance-rejection and family functioning are incorporated as functions of beliefs and
behaviors measured include those associated with food and eating.

Freedman Melson (1980) supports the notion that environmental information is
never neutral, but rather loaded with symbolic meanings of past history and present
experience. These symbolic meanings invite symbolic interactionism into the framework
of this study in an effort to clarify how situations are deﬁned and, in turn, inﬂuence
individual systems embedded in relational contexts.

Symbolic Interactionism

Ernest Burgess called upon his colleagues in 1926 to study the family as a “living,
changing, growing thing,” contending that “the actual unity of family life has its
existence... in the interaction of its members” (Burgess, 1926, p. 5). Since Burgess’s
plea, proponents of symbolic interactionism for the study of families acknowledge the
limitation of other theories that aim to discuss social factors as independent variables and
behavioral variables as dependent (Burr, Leigh, Day, & Constantine, 1979). Deﬁnitional
attributions, according to symbolic interactionism theory, are the meaning-making
processes that individuals engage in whereby their perceptions are shaped and molded.
For instance, a child may attribute 3 parents’ behavior of continually leaving her with a
neighbor as rejecting. It is this interpretation that is suggested in the language of
deﬁnitional processes or attributions. Some theorists suggest that more variance in the
dependent variables may be accounted for by focusing on these deﬁnitional processes as
more direct causes (Burr, et al., 1979). The current study utilizes this line of thinking and
aims to look at the deﬁnitional attributions as predictive of behavioral sequences,

speciﬁcally disordered eating.

10

Symbolic interactionism has a number of contributors to its multi-faceted and
complex viewpoint of studying families. George Herbert Mead, Charles Horton Cooley,
William Isaac Thomas, and Herbert Blumer are a few whose ideas are most clearly
articulated in the literature. Cooley, in particular, laid a groundwork within symbolic
interactionism to make it highly compatible with Bronfenbrenner’s human ecology
theory. Cooley has acknowledged that the individual and society are two sides of the
same coin with no individual existing apart from society (Longmore, 1998). In
accordance with human ecology theory, symbolic interactionism proposes families to be
social groups wherein members develop their identities and self concept through social
interaction (LaRossa & Reitzes, 1993). Meanings emerge from interaction between
subject and object (LaRossa & Reitzes, 1993). These meanings elicit congruent behaviors
as a result of the deﬁnition given by the interpreter. Symbolic interactionism is a solid
foundation for the proposed study in that it has a micro orientation, focusing on intra- and
interpersonal phenomena (Burr, et al.,

1979). Symbolic interactionism adds a dimension to human ecology theory by
acknowledging interpretations and perceptions of individuals, whereas human ecology
theory emphasizes the environment in which these interpretations are made.

According to symbolic interactionism, people deﬁne situations based on their own
personal experiences and sense of self. Mead believed that people learn about themselves
through interactions with others, therefore making interactions necessary for the full
development of a sense of self (Mead 1934/1956). Mead is known for his contributions to
the theory including the conceptualizations of self and mind (W inton, 1995). The self is

said to refer to an individual’s capacity to step outside and view themselves as part of the

11

environment. George Herbert Mead held an optimistic view that when individuals take
appraisals from others, they are interpreted through a selective process that ﬁlters out the
negative aspects of self, holding onto only those that are positive. Mind, Mead’s second
primary concept, postulates that a sequence develops from thinking to action. Cast (2003)
highlights that meaning is attached to one’s identity, leading to particular behaviors, such
as disordered eating. For instance, if one perceives herself as rejected, resulting behaviors
might be aligned with this identiﬁcation to make it more valid. Alternatively, attempts at
remedying the rejection, such as trying to be more perfect and lovable may be another
way for the child to respond.

Mead also postulated that thinking, the primary activity of the mind, is a
necessary precursor to action (Winton, 1995). In applying this concept of thoughts
precede action in the current study, a young woman perceiving herself as rejected by her
parents, will feed into actions, likely aimed at conﬁrming or denying the perception. The
relationship being discussed here is not linear when symbolic interactionism is combined
with human ecology theory whereby systems affect systems. The way in which people
deﬁne, or perceive a situation will affectthe action they take in the context of that
situation (W inton, 1995). Further, individuals’ perceptions of others’ thoughts and
feelings about them may be internalized and signiﬁcantly inﬂuence their construction of
the self-concept (W onderlich & Klein, 1996).

In contrast to Mead’s conceptualization of the self as being highly responsible and
able to ﬁlter through messages, discarding the negative or harmful and choosing to
incorporate the positive, Cooley presents a somewhat different interpretation. Commonly

known as the “looking glass self”, Cooley believed an emergent image of self is

12

internalized based on the input received from others (Longmore, 1998). The following
quote comes from Murdock (p. 108, 1992) and highlights the concept of the looking glass
self as well as the importance of paternal inﬂuence on female development to be
discussed in Chapter 2,

“A young girl’s relationship with her father helps her to see the world through his

eyes and to see herself reﬂected by him. As she sees his approval and acceptance,

she measures her own competence, intelligence, and self-worth in relation to him
and to other men.”
Just as a girl or woman’s sense of approval and acceptance from her father can positively
impact her development and adjustment, a sense of rejection from her father will likely
negatively inﬂuence her growth.

Cooley further describes that a person takes on the identity of what she believes
others see, in essence, the expectations of others are central to the development of self
perceptions (Winton, 1995; Longmore, 1998). The looking glass self is thought to emerge
in the context of primary groups, with family being one example. Cooley acknowledges
in his writing, “it is in the family that an infant becomes aware of others and interested in
gaining their approval and support for a positive self-conceptualization” (LaRossa &
Reitzes, 1993). The family has been said to be the major institution of socialization,
thereby serving as the foundation for the development of self for each of its members.

A number of assumptions exist within the symbolic interactionism framework that
serve as a guide and foundation for the proposed study. These assumptions include: 1)
perceptions are one important determinant of how a situation is handled, 2) people react
to something according to it’s idiosyncratic meaning, or what it symbolizes to them, 3)

meaning is derived through interaction with others, 4) interpretation emerges about what

a person learns as they comes into contact with different situations and experiences, 5)

l3

humans develop ideas about themselves through interacting with others, 6) humans are
reﬂexive whereby they incorporate experiences into a guide for future behavior and
ﬁnally, 7) individuals are inﬂuenced by family as well as the larger culture and society
(Ingoldsby, Smith, & Miller, 2004; Burr, et al., 1979).

William Isaac Thomas’s contribution to symbolic interactionism included the
creation of a concept known widely as “deﬁning the situation” (LaRossa & Reitzes,
1993). Identities both result from, and are fostered by interactions according to the
process of “deﬁning the situation” (Cast, 2003; Burr, Leigh, Day, & Constantine, 1979).
Subjective meaning is given to a particular situation by an individual (Burr, et al., 1979;
LaRossa & Reitzes, 1993). Deﬁnitions therefore are synonymous with the concept of
perceptions when using symbolic interactionism language (Burr, et al., 1979). Thomas
described the phrase “deﬁnition of the situation” to mean one cannot understand behavior
without also understanding the subjective perspectives of the individuals involved in the
situation (Ingoldsby, Smith, & Miller, 2004). Tenets of postmodernism and social
constructionism can be clearly be linked to Thomas’s contributions to symbolic
interactionism, whereby reality is thought to be constructed through a process of
interaction in groups (Hoffman, 1990). Thomas theorem states, “if an individual deﬁnes a
situation as real, they are real in their consequences” (Burr et al., 1979; LaRossa &
Reitzes, 1993). At its core, the Thomas theorem contends that social situations are not
completely determined by objective conditions, but are also inﬂuenced by attitudes and
subjective deﬁnitions of the situations held by the interacting subjects (LaRossa &

Reitzes, 1993). In other words, whatever the deﬁnition is according to the individual, this

14

deﬁnition will inﬂuence how he or she responds through action (LaRossa & Reitzes,
1993).

Symbols are products of social interaction, with their meaning ascribed by the
way we see others using them (LaRossa & Reitzes, 1993). Symbols are deﬁned based on
the context of the situation, being something learned from interacting with others in an
environment (Ingoldsby, Smith, & Miller, 2004). Additionally, interdependence exists
between symbols and interaction. Interaction involves communication between at least
two people where reactions and modiﬁcations of behavior take place (Ingoldsby, Smith,
& Miller, 2004). In the context of the proposed study, the interactions that take place
between parents and child over time serve to create meaning for the child about how his
or her parents respond to him or her. The symbol, in the conceptualization of this study,
includes parental acceptance or rejection as evidenced by various aspects of the

interaction from verbal to non-verbal exchanges.

15

Theoretical Map

__.___

fp—

 

 

\
/// \\\\
// Familial Microsystem Environment \\\\
// /”_\\.\ FEEDBACK LOOPS // \ \\\
/ //
, FATHER \V MOTHER\ \_ \\
<90EPINITION DEFINITION
PERCEPTION 01 THE PERCEPTIONQ 01' THE
SITUATION SITUATION
/
INTERPRETATION/ mmmmw //
émcno /
\\z><<‘//\ /
/
INTERPRETATION
\ mm 1:10: /
/
\ PROCESSES DEFINITION\}\ 3m
\ PERCEPTION Q of, T1111
\ SITUATION /}/
\\‘ ’,//
\\ x \ \\\ // ,//
\\ \\ CHILD // ///
I\\\ \\ ,// ///
\\\\\——.—/// //

Figure 1: Theoretical Map
Research Questions

The research questions for this study examined a predictive relationship between
measures speciﬁc to perceptions of parental acceptance-rejection and family functioning
with disordered eating symptomatology. The overarching research question proposed in
the study asked:

0 Are parental rejection and family functioning related to disordered eating

symptomatology?
In addition to this core research question, several others have been developed to better

understand the variables’ relationships to one another.

16

. Does parental acceptance-rejection differ between individuals with disordered

eating and those without?
Does family functioning differ between individuals with disordered eating and
those without?
Does the presence of parental rejection predict disordered eating
symptomatology?
Does the perception of unhealthy family functioning predict disordered eating
symptomatology?
Is maternal or paternal rejection a better predictor of disordered eating
symptomatology?

Hypotheses

Disordered eating may be one form of coping behavior employed to reduce or

minimize some of the psychological pain produced by perceived parental rejection and a

negatively functioning family. Each hypothesis below corresponds with the research

question identiﬁed above. This study has been designed to address the following

hypotheses:

1.

Individuals presenting as disordered in their eating will perceive their parents as
more

rejecting than individuals without disordered eating symptoms.

Individuals presenting as disordered in their eating attitudes and behaviors will
perceive their families to be less healthy in terms of family functioning, than

individuals without disordered eating difﬁculties.

17

3. Individuals indicating more parental rejection will exhibit greater degrees of
disordered eating.

4. Individuals indicating unhealthy family functioning will exhibit greater degrees of
disordered eating.

5. Although it is hypothesized that individuals reporting higher levels of disordered
eating will perceive both their mothers and father to be more rejecting than the
non-eating disordered group (hypothesis 1), it is further hypothesized that
individuals reporting disordered eating symptomatology will perceive their fathers
to be more rejecting than their mothers.

Conceptual Model

  
    
   
   
 
       
   
  
   
     
    

ERCEPT IO
OF PATERNAL
ACCEPTANCE-

REJECTION

  

 

 

PERCEIVED
UNHEALTHY
FAMILY
FUNCTIONING

DISORDERED
EATING

 

 

 

 

PERCEPTION OF
MATERNAL
ACCEPTANCE-
REJECTION

Figure 2: Conceptual Model

18

Methodology

Cross-sectional data was collected from a sample of college-students attending a
large Mid-western university to facilitate a correlational design to analyze data. Purposive
sampling procedures were used to assist in attaining a sample representative of
undergraduate females between the ages Of 18 and 25 years. Students enrolled in large
undergraduate psychology and family and child ecology courses were invited to
participate in the study by completing a battery of assessments via computer or classroom
recruitment. Self-report instruments aim to assess perceptions of parental rejection-
acceptance, family functioning, and disordered eating. Data were analyzed by using t-
tests for research questions 1 and 2 and multiple regression analysis for questions 3, 4,
and 5.

Overview of the Dissertation

Chapter 1 has described the purpose, guiding theoretical frameworks, research
questions, hypotheses, conceptual model, and provided a brief overview of the project’s
methodology. Chapter 2 provides a review of the relevant literature pertaining to
disordered eating in the family context. In addition, the literature review discusses the
construct of rejection and Ronald Rohner’s conceptualization of the parental acceptance-
rejection framework. Chapter 3 delineates the speciﬁc methodology that will be used to
address the study’s research questions. Chapter 4 presents the statistical results. Finally,
Chapter 5 will provide a discussion based on the data as well as limitations, clinical

implications, and future directions.

19

CHAPTER 2
Review of the Literature
Introduction
The previous chapter highlighted the purpose of the present study to ﬁll a gap in
research that addresses the relationship between parental acceptance-rejection and family
functioning as predictive of disordered eating in females. In this chapter, existing
literature will be reviewed to support such a study, including familial inﬂuence on
disordered eating, perception of family functioning, and the construct of parental
rejection. The present study is considered exploratory in nature because no other study to
date has explored these independent variables of parental rejection and family
functioning as related to disordered eating attitudes and behaviors representative of
anorexia nervosa and bulimia nervosa symptomatology. Although the literature addresses
other ecological factors contributing to the existence of disordered eating such as peers,
genetics, media inﬂuence, cultural expectations, and developmental life cycle stage, this
study is aimed at focusing on family inﬂuences, in particular, how the perception of
maternal and paternal characteristics impact the presence of symptoms. Therefore,
according to Human Ecology Theory, the microsystem chosen for analysis in this study is
that Of the immediate family with consideration that others exist (university, peers, etc.),
but are not being measured at this juncture.
Disordered Eating and Familial Environment
Early Family Typologies
Since the earliest case descriptions of anorexia nervosa, dating back to 1860,

families have been reported as critical to the etiology and maintenance of the. disorder

20

(Ward, Ramsay, Tumbull, Benedettini, & Treasure, 2000). Family clinicians and
theorists, beginning with Salvador Minuchin and Mara Selvini-Palazzoli, have identiﬁed
the familial environment with an eating disordered member to include particular
characteristics. Minuchin and colleagues (1978) contend, through clinical observations
that four main deﬁning features exist within families with eating disordered members.
These four features include enmeshment, rigidity, over protectiveness, and lack of
conﬂict resolution (Minuchin, Rosman, & Baker, 1978; Minuchin & Fishman, 1979).
Selvini Palazzoli (1978) ﬁnds many similar family traits in her practice, but uses different
language when communicating them. Selvini Palazzoli ﬁnds the family characteristics of
eating disordered environments as including a high degree of marital dysfunction,
leadership problems, rejection of communicated messages, poor conﬂict resolution,
covert alliances or “denied coalitions”, blame shifting, and extreme rigidity
(Vanderlinden & Vandereycken, 1989; Selvini Palazzoli, 1978). Empirical studies have
generally upheld that anorexic families tend to display greater boundary pathology, in
particular enmeshment patterns as originally postulated by Minuchin and bulimic families
evidencing heightened levels of hostility and conﬂict as well as nurturance deﬁcits (Kog
& Vandereycken, 1989; Humphrey, 1986; Johnson & Flach, 1985; Ordman &
Kirschenbaum, 1986)
Content versus Process Driven Research

Most studies focusing on family in regards to eating disorders have looked at
areas such as family and parental functioning, communication patterns (White, 2000),
parental attitudes toward weight and body (White, 2000), parental modeling (MacBrayer,

Smith, McCarthy, Demos, & Simmons, 2001; Kichler & Crowther, 2001) intra-familial

21

teasing (MacBrayer et al., 2001; Kanakis & Thelen, 1995), and parents own issues with
weight, food, and their bodies (Steiger, Stotland, Trottier, & Ghadirian, 1996; Steinberg
& Phares, 2001).

Research relating the content associated with eating disorders such as food,
weight, appearance-driven comments, and body image has a large representation in
academic journals. MacBrayer, Smith, McCarthy, Demos, & Simmons (2001) developed
a measure of family of origin food-related experiences and family modeling inﬂuences
called the Family History Inventory (FHI). The FHI contains 14 primary scales including
the assessment of such factors as teasing by various family members (siblings, father,
grandparents, mothers) relationship with food, rules pertaining to eating, meal structure,
and the use Of food as rewards within the family system. Results from their study
indicated higher rates of bulimic symptoms were correlated with teasing by one’s family
about one’s weight and observation of negative maternal modeling with respect to food
(MacBrayer et al., 2001).

Another form of content-based research includes teasing and criticism with
regards to the topic of weight and appearance in peer groups versus families. Kanakis and
Thelen (1995) found bulimics and sub.clinical bulimics reported being affected more by
family criticism and teasing on the topic of weight compared to peers. Speciﬁcally, there
were no group differences in the effect of being teased by one’s peers in the three groups
studied: bulimic, sub-clinical bulimic, and non-eating disordered (Kanakis & Thelen,
1995).

For instance, White (2000) found that parental comments inﬂuence children’s

degree of body dissatisfaction, fears about being fat, and weight-loss attempts. Kanakis

22

and Thelen (1995) add that individuals with bulimia and sub-clinical features Of the
disorder reported greater effects from family teasing than by peer teasing about being
overweight as compared to controls. More recently, Haworth-Hoeppner (2000) identiﬁed
four conditions that foster the development of an eating disorder in families to include a
critical family environment, coercive parental control, unloving parent-child relationship,
and a main discourse on weight. The factors identiﬁed above including family teasing,
negative comments about children’s bodies, weight, and shape, critical and controlling
parenting, and a family environment focused on weight and appearance can all be said to
contribute to an overt form of rejection regardless of the child’s interpretations. The
current study is more interested in if a child perceives herself as rejected outside of the
realm of weight, shape, and appearance-based factors, a more generalized rejection. The
key piece to measuring this construct of rejection is that it is indicated by the child,
involving the perceptual component.

Beyond research focusing on symptom driven, or content-oriented studies, there is
an opportunity for process research to emphasize more on the internal workings of the
family as related to the etiology and maintenance of disordered eating. Process-oriented
research would also be interested in the perceptions of family members regarding
functioning and general family environment and apply these perceptions to the presence
or absence of symptoms. Research targeting less content driven analysis such as dieting,
and more emphasis on the process of family interactions has included such variables as
parenting style, family modeling, intergenerational transmission, and overall family
functioning of families with and without disordered eating. One study focusing on these

family process variables in participants with binge eating behaviors, found these

23

individuals tend to rate their families as having more conﬂict and control than a group of
controls (Hodges, Cochrange, & Brewerton, 1998). Fonseca, Ireland, and Resnick (2002)
also found their index group of females to report signiﬁcantly lower levels of family
connectedness and communication using a non-clinical sample of high school students.
Another study has found that families including a bulimic individual tend to exhibit lower
levels of family support and connectedness (Fairbum, Welch, & Doll, 1997). Females
struggling with bulimia report experiencing their families as distant, conﬂictual, and non-
supportive (Humphrey, 1989; Johnson & Flach, 1985). A presumably fertile environment
for perceived rejection has been evidenced in the studies just cited, but thus far questions
pertaining to their feelings of parental acceptance versus rejection have not been asked Of
study participants by researchers. Again, an environment that is distant, conﬂictual, and
non—supportive may foster the likelihood that a child feel rejected, but does not guarantee
a child will feel rejected. This study looks to uncover if the children who both report
these types of environments as well as identify themselves as rejected have a higher
likelihood of displaying disordered eating symptomatology.

Perceived Family Functioning and Environment in Eating Disordered Families

As stated earlier, a speciﬁc kind of process-oriented research can be identiﬁed as

including individual’s perceptions of family functioning and overall environment.
Neumark-Sztainer, Story, Hannan, Beuhring, and Resnick (2000) surveyed a sample of
9,943 seventh, ninth, and eleventh grade students to ﬁnd that youth at increased risk for
disordered eating included those who perceived family communication, parental caring,
and parental expectations as low. Several researchers (Johnson & Flach, 1985; Ordman &

Kirschenbaum, 1986; Humphrey, 1988; Kog & Vandereycken, 1989) have studied the

24

perceptions of individuals with bulimia on the area of family functioning and parental
characteristics. These studies have yielded perceptions of family functioning in
disordered eating groups as compared to control groups as being less expressive, less
cohesive, more conﬂictual, disorganized, achievement oriented, and hostile. Further,
parents of individuals with bulimia were perceived as being more blaming, rejecting,
neglectful and less nurturant and comforting compared with controls.

Felker and Stivers (1994) examined the relationship between family environment
and the risk of developing an eating disorder using the Family Environment Scale. A
signiﬁcant relationship was found between family environment and the development of
anorexia or bulimia in an adolescent population. Speciﬁcally, participants who perceived
their family environments to be less cohesive, less organized, less independent, less
expressive, more conﬂictual, higher on parental control, and including a greater
orientation toward achievement were associated with an increased risk for the
development of an eating disorder (Felker & Stivers, 1994). Another study of interest to
utilize Moos and Moos’s Family Environment Scale targeted 175 female undergraduates
(Scalf—Mclver & Thompson, 1989). This study investigated the relationship between
degree of bulimic symptomatology and family dynamics. Results indicated a negative
correlation between cohesion and bulimia symptoms, with commitment and support
decreasing as bulimic symptoms increase.

In a study looking at the link between attachment style and family functioning as
related to the presence of eating disorders, families Of eating disordered patients were
found to be less cohesive, expressive, and encouraging Of personal growth as compared

to controls (Latzer, Hochdorf, Bachar, & Canetti, 2002). Using an attachment theoretical

25

basis, these researchers hypothesized that a family environment unable to provide a sense
of security and availability contributes to the presence of detachment that characterizes
mental disorders in general, and eating disorders in particular (Latzer, et al., 2002). Using
the Adult Attachment Scale and the Family Environment Scale, Latzer and colleagues
(2002) discovered an association between both an insecure attachment and particular
family environment characteristics with eating disorders. Speciﬁcally, lower levels of
cohesiveness were found in families with a bulimic member and lower levels of
expressiveness were found in both anorexic and bulimic families (Latzer, et al., 2002). It
is important to note that the differences in family environment disappeared when
attachment style was controlled in this study. The researchers interpret this ﬁnding to
mean that attachment style may be a primary differentiating factor in the onset of eating
disorders, however, they go on to postulate that the interaction between attachment and
environment is what determines the development of an eating disorder.

The present study has utilized the Family Assessment Device (Epstein, Baldwin,
& Bishop, 1983) to measure family functioning. The following studies have also included
this measurement with eating disorder populations. McGrane and Carr (2002) have
studied the link between Eating Disorder Inventory-2 scores indicating more severe
levels of psychopathology, with scores on the Family Assessment Device (FAD)
indicating more perceived family dysfunction. A group of 27 individuals scoring above
the clinical cut-off on the EDI-2 were compared to a group of 27 controls. T-tests were
used to assess differences between the groups with results indicating signiﬁcant

differences (p<.01) on the dimensions of problem solving, roles, affective responsiveness,

26

and general functioning, with women in the eating disordered group reporting greater
difﬁculties in these family areas.

Compared with controls, families with an eating disordered member report greater
dysfunction on the communication and affective responsiveness sub-scales of the Family
Assessment Device (Steiger, Liquomik, Chapman, & Hussein, 1991). Additionally,
women having both anorexia and bulimia have been found to rate their families as more
unhealthy on the affective involvement and behavior control scales of the Family
Assessment Device (Waller, Calam, & Slade, 1989). Women with bulimia were also
shown to rate their family as having problem solving difﬁculties according to the FAD
(Waller, Calam, & Slade, 1989).

Waller, Slade and Calam (1990) compared perceptions of family functioning with
three groups including anorexic (n=14), bulimic (n=34), and non-eating disordered
women (n=30). The researchers used the Family Assessment Device which measures
problem solving, roles, affective responsiveness, affective involvement, communication,
behavior control, and general functioning. Results indicate that the anorexic and bulimic
groups did not differ signiﬁcantly, therefore providing a rationale for coming these two
groups. The combination of these two symptom groups will be replicated in the present
study through the use of the Eating Attitudes Test—26. Wonderlich and Swift (1990) also
found a lack of difference among various eating-disorder subtypes with regard to their
perceptions of their parental relationships.

Waller (1994) conducted a study looking at the perceptions of family functioning
from a sample of 81 bulimic women to explore if a relationship existed with the

dependent variable, eating attitudes and behaviors. Multiple regression analyses for the

27

sample of 81 women indicated a non—signiﬁcant predictive relationship between the
Family Assessment Device and the dependent variable derived from the Eating Attitude
Test score. Waller (1994) looked more closely at speciﬁc symptomatology, however, to
discover that increased bingeing behaviors were related to increased perception of poor
problem solving and low degrees of cohesion within the family. Additionally, women
rating their families as more cohesive binged less frequently. Extrapolating from these
ﬁndings, Waller (1994) suggests the possibility of protective and risk factors associated
with speciﬁc symptoms. For instance, perception of poor problem solving served as a risk
factor for bingeing with perception of higher levels of family cohesion serving as a
protective factor for bingeing (Waller, 1994). Risk factors have been identiﬁed as
increasing the likelihood of maladaptation (Rutter, 1990) whereby compensatory factors
are deﬁned as reducing the risk of developing problems such as emotional, psychological,
and social (Steinhausen & Winkler Metzke, 2001).

In a study looking to identify factors that inﬂuence emotional and behavioral
abnormalities in adolescence, perceptual components of parental acceptance-rejection
were implicated as both risk and compensatory factors for mental health measures.
Speciﬁcally, using analysis of variance, parental acceptance was shown to have a
signiﬁcant main effect as a protective factor and parental rejection was indicated as a risk
factor for both internalizing and externalizing scores (Steinhausen & Winkler Metzke,
2001). Another study indicating the conceptualization of parental acceptance as
protective or compensatory and rejection as risk, evidenced that parental acceptance is
associated with less substance use and abuse by youth (Campo & Rohner, 1992). These

ﬁndings serve as the basis for the proposed study, as parental rejection is being

28

hypothesized as a risk factor, therefore contributing to the presence of more disturbed
eating attitudes and behaviors.
Parental Inﬂuences

A speciﬁc dyadic relationship that has been extensively studied in the literature is
that between an eating-disordered daughter and her mother. Dating back to the writing
and clinical work of Hilde Bruch (1974), the mother and eating disordered-daughter
relationship has been examined through a lens of attachment, whereby the mother
intrusively superimposes her own needs onto the child, resulting in the child not being
able to differentiate her own impulses (Ward, et al., 2000). Mother-daughter dyads were
also investigated by Rupp and Jurkovic (1996), looking at individuation and perspective-
taking differences between bulimic and non-bulimic female adolescents and their
mothers. Further, daughters’ perceptions of family approval of her appearance were
important predictors of higher body esteem. Mothers’ own reports of their approval of
their daughters’ appearance were not signiﬁcantly related to daughters’ body esteem
(McKinley, 1999).

Phares (1992) found a tendency in the literature to blame mothers for the vast
majority of their children’s psychological problems. A review of clinical child and
adolescent research from 1984 to 1991 revealed that 48% of studies looked exclusively at
maternal factors, whereas only 1% involved fathers exclusively (Phares, 1992). A gender
variation in parental inﬂuence has long been speculated in the literature with eating
disorders, often implicating maternal inﬂuences in the development of eating disorders
(Kichler & Crowther, 2001; Moreno & Thelen, 1993). For instance, highly appearance-

invested mothers who value dieting behavior, strive for weight loss, or engender family

29

competition based on physical attractiveness have been found to promote the
development of a negative body image in their daughters (Haworth-Hoeppner, 2001).
Another study by Moreno and Thelen (1993) looked at the difference between mothers
and fathers of three groups including bulimic, sub-clinical, and normal. These
investigators concluded mothers of bulimic individuals, and in some cases individuals
reaching sub-clinical levels of pathology, were more likely to restrict their daughters’
food intake and encourage their daughters’ to diet and exercise in order to lose weight.
The fathers in the study showed no signiﬁcant differences between the three groups,
indicating less investment in their daughter’s weight issues (Moreno & Thelen, 1993;
Kanakis & Thelen, 1995). In a follow-up study, bulimic and subclinical females were
found to perceive more pressure to lose weight from their mothers than their father
(Kanakis & Thelen, 1995).

Through an analysis of over 150 studies researching psychosomatic disorders, a
picture of maternal characteristics correlated with the presence of a psychosomatic-type
disorder has been drawn to include domineering, overly involved, demanding,
unempathic, insensitive, and rejecting, either overtly or covertly (Sackin, 1985). The vast
majority of data presented in this review focuses on the maternal inﬂuence, all but
ignoring the father’s involvement in the development or maintenance of psychosomatic
disorders. In an effort to remedy this limitation, the research being proposed will gather
data using the launched child’s perception of both her parent’s acceptance or rejection.

Academicians and researchers have been much less interested in the father’s role
and relationship with his daughter as relating to the presence of disturbed eating attitudes

and behaviors. An exception has been writer Margo Maine, author of “Father Hunger”

30

(1991). Maine postulates that females in particular exhibiting eating disorders experience
an unﬁlled longing for a father’s presence, both behaviorally and emotionally, in their
lives (Maine, 1991). Unfortunately, Maine did not empirically test her contention using
research, but only using clinical experiences as evidence.

The literature that does exist regarding fathers of eating disordered females
suggest that they are typically considered to be “distant”. In his remoteness, the distant
father of an eating disordered daughter tends not to be involved and fails to understand
her needs for love, encouragement, and affection (Friedman, 1997). In general, these
fathers are written about as psychologically unavailable to their daughters (Secunda,
1992).

A recent study conducted by Dominy, Johnson, and Koch (2000) found that
women with binge eating disorder reported their father to be more rejecting and less
warm in comparison to mothers. This ﬁnding supports an analysis which looks at the
perceptions of both maternal and paternal acceptance-rejection levels rather than a
uniﬁed parental measure. Wonderlich, Ukestad, and Perzacki (1994) also found results
indicating preliminary indications to suggest that a daughter’s perceptions of her father
may play a role in the development of eating disorders.

Wonderlich, Klein, and Council (1996) set out to examine how perceptions of
parental behaviors such as attack, hostile disengagement, control, and submission
speciﬁcally relate to the bulimic individual’s self-attach and self-restraint. This team
found fathers to be perceived by their bulimic daughters as signiﬁcantly less friendly than
their mothers. Whereby relationships with mothers were reported to consist of less

interpersonal engagement and more mutual hostile withdrawal as compared to controls.

31

Based on this more recent research that focuses more on the process of
perception rather than on appearance or dieting variables (often present in maternal-
focused literature), this study hypothesized that perceived paternal rejection serves as a
better predictor than maternal rejection for the presence of disordered eating. T o clarify,
it was hypothesized that both paternal and maternal rejection were signiﬁcant in
predicting disordered eating, but more speciﬁcally, it is thought, based on the evidence
presented here, that paternal rejection accounts for more total variance than maternal
rejection.

One study has used both paternal and maternal acceptance and rejection
measurements with African American and European American youth, in relation it to
psychological adjustment (Veneziano, 2000). The investigator was interested speciﬁcally
in the variance accounted for by paternal versus maternal acceptance in psychological
adjustment based on recent research suggesting that paternal warmth is often more
signiﬁcantly related to children’s development than maternal warmth particularly in
European American families. Using multiple regression analyses, it was found that
youths’ self-reported psychological adjustment scores using the psychological adjustment
questionnaire, PAQ, was signiﬁcantly related to perceptions of both paternal (1:.53,
p<.001) and maternal acceptance (F46, p<.001). Interestingly, in the European
American sample, maternal acceptance-rejection as a main effect was eliminated from the
model due to a lack of statistical support. This ﬁnding signiﬁes the importance of
measuring paternal acceptance-rejection, which has not consistently been included by
researchers as Ronald Rohner’s initial assessment was created only for mothers.

Although this study is not speciﬁc to eating attitudes and behaviors, it highlights the

32

importance of considering both parents’ inﬂuence as opposed to measuring the parental
collective or only measuring one parent, which traditionally has been the mother.
Construct of Rejection

In an extensive review of the literature completed by Rohner and Nielsen (1978)
over two decades ago, the researchers documented 600 studies completed with an
incorporation of parental acceptance-rejection language. Unfortunately, many of these
studies cannot be compared due to inconsistent Operationalization of parental rejection.

Early studies in the area of parental rejection tended to focus on maternal
deprivation, whereby Bowlby described the child lacking warmth, intimacy and a
continuous relationship with his or her mother (Bowlby, 1966). Three ways of measuring
parental rejection including monitoring parental behaviors and coding them as rejecting,
accepting, or neutral, inquiring about parental feelings and attitudes toward their
child(ren), or assessing the child’s perceptions of his or her parent’s rejection or
acceptance (Envoy, 1981). It is this latter method of measurement and conceptualization
of the concept that is of particular interest for the current study. Embedded within a
child’s perception lies the concepts of interpretations and meaning, whereby the
subjective experience of rejection is critical to its presence rather than Objective
observations (Hawkes, 1957).

John Envoy incorporated this meaning into his phenomenological work, studying
individuals who presented as feeling rejected by his or her parents, regardless of the
extent to which the presented feelings corresponded to parental attitudes or behaviors
(Envoy, 1981). Further, in a study measuring the intergenerational continuity Of parental

rejection and its association with depression, the researchers operationalized rejection

33

similarly to both Envoy and Rohner, stating it is a felt lack of parental warmth and caring
with parents being emotionally withdrawn and expressing dissatisfaction to their children
(Whitbeck, Hoyt, Simmons, Conger, Elder, Lorenz, & Huck, 1992). Barnow, Schuckit,
Lucht, John, and Freyberger (2002) included parental rejection as a factor in a path
analysis researching its effects on alcohol problems in teenagers. These researchers
utilized a similar deﬁnition of the construct including lack of emotional warmth or
support as measured by the EMBU, a Swedish measure (Barnow, et al., 2002).
Unfortunately, neither Envoy (1981) or Whitbeck, et a1. (1992) used a standardized
measurement of parental rejection to make comparisons to their ﬁndings suitable to
future research. The incorporation of Rohner’s universal parental acceptance-rejection
questionnaire, a standardized measure, is proposed for utilization in the current study for
this reason of future replication and ease of comparative analysis.

Ronald Rohner made great strides to operationalize the abstract construct of
parental rejection and acceptance by placing the two concepts on Opposite poles of a
continuum he describes as a warmth dimension. Another writer, John Envoy, was also
been intrigued by the parental rejection construct as it related itself to therapeutic work
with clients over his career as a clinician. Clients have reportedly described rejection as
“knowledge that they were not loved and wanted by one or both parents” (Envoy, 1981).

Parental rejection may be assessed either by the subjective experience of the child
or parent or by external measurement by observers looking in on the parent-child
interactions. The focus of the proposed study aims to measure the construct of parental
acceptance-rejection from the viewpoint Of the child. This strategy was chosen based on

the possibility that a child can experience rejection or feel unloved without consistent

34

parent reports or observation evidence. Rohner accounts for this idea in PART by the
inclusion of undifferentiated rejection. Within this construct, an individual perceives
themselves to “be unloved without necessarily feeling that their parents are either overtly
hostile/aggressive toward them, or actively indifferent/neglecting” (Rohner, 1991, p. 2).
Further, according to PART, parental rejection has most consistent and predictable
effects on children where they perceive themselves as being rejected by their parent
(Rohner, 1991).

Webster and Palmer (2000) conducted a study aimed at looking at the family
background of women with bulimia nervosa (BN), anorexia nervosa (AN), both AN and
BN (MIXED), depression (DEP), and controls (CON). The Childhood Experience of
Care and Abuse Interview (CECA) was used to assess family background. Although the
terminology differs, a rejection construct is implied in the measure’s assessment. For
instance, the CECA is composed of 5 scales which include two quite similar to Rohner’s
construct of rejection. The parental indifference scale reﬂects general neglect and a lack
of interest and attention from the parents. A second scale, antipathy, is deﬁned as a
negative quality present in the parent-child relationship as subjectively experienced by
the child such as dislike, coldness, or hatred. Both of these scales coincide with Rohner’s
conceptualization which includes the scales of aggression/hostility, neglect/indifference,
and rejection to measure perceived rejection from parents. In the Webster and Palmer
(2000) study it was found that the indifference variable described above was statistically
signiﬁcant for the BN and MD(ED groups and the antipathy variable was statistically
signiﬁcant for the MD(ED group. Thirty-four percent of the bulimia participants and 40%

of the MD(ED group identiﬁed indifference as present in their parent-child relationship

35

growing up. Fifty percent of the MIXED group identiﬁed antipathy to be part of this
relationship.

A study comparing the early family experiences of women with bulimia nervosa
(n=30), women with major depression (n=15), and a group of normal controls (n=100)
found bulimic women perceived both parents as more rejecting in childhood as compared
to both major depressive and normal females (Studart, Laraia, Ballenger and Lydiard,
1990). Additionally, Dolan, Lieberman, Evans, and Lacey (1990) compared perceptions
of family interactions between bulimic females (n=38) and a group of normal controls
(n=40). Overall bulimic women were found to report a poorer relationship with their
parents, perceiving their parents as inattentive and uninvolved. This perception may be
incorporated into the rejection construct described by Rohner to be studied in the current
project. Rohner’s construction of rejection is made up of four factors including
neglect/indifference, aggression/hostility, undifferentiated rejection (subjective feeling of
rejection), and lack of warmth/affection.

The idea or construct of rejection is embedded into a number of instruments and
studies, not always referred to as “rejection”. Another example of this differential
nomenclature is present within the Parental Bonding Instrument (PBI). The PBI aims to
operationalize the relationship between parent and child as does Rohner, but it chooses
different language and theory (attachment rather than socialization) to do so.

Parent Acceptance-Rejection Framework

Ronald Rohner developed a framework based on socialization theory, which he

has called Parental Acceptance-Rejection Theory (PARTheory hereafter). PARTheory

attempts to predict and explain signiﬁcant antecedents, consequences of perceptions of

36

parental acceptance and rejection (Rohner, 1986). The notion of parental rejection and its
hypothesized consequences have been studied throughout the twentieth century (Rohner
& Nielsen, 1978).

Rohner has conceptualized a theory, which he has called PARTheory to capture
the constructs of acceptance and rejection, which lie on the continuum of warmth. The
continuum of warmth includes affection, support, caring, nurturance, and love, or lack of
these features. Rejection (lack of parental warmth & affection) would lie on one end and
acceptance on the other. Acceptance has been characterized as physical and verbal
expressions of love from parents to child and rejection as dislike, disapproval,
resentment, or indifference from parents toward children (Rohner 1986; 1991).

PART is a theory of socialization attempting to explain and predict consequences
of both parental acceptance and rejection for development in children and adults (Rohner,
1991). Rohner and Britner (2002) found parental rejection tends to be an “excellent
predictor of psychological and behavioral problems” by way of a meta-analysis.
Conceptualized in four ways, parental rejection may include: 1) cold and unaffectionate,
2) hostile and aggressive, 3) indifferent and neglecting, and 4) undifferentiated rejecting
(Rohner & Britner, 2002). The fourth category of “undifferentiated rejecting” is of
particular interest as it refers to the perceptual piece, or feeling unloved without objective
indicators (Rohner & Britner, 2002).

Rejection is manifested by parents in two primary ways, including
hostility/aggression or indifference/neglect (Rohner, 1991). These manifestations include
feelings, attitudes, or behaviors exhibited by parents toward their child(ren). Hostility, for

instance, refers to feelings of anger or resentment toward the child. Alternatively,

37

indifference refers to a lack of concern or interest in a child, including ignoring children’s
bids for attention or remain remote or inaccessible to them (Rohner, 1991). PART states
that either of these forms of parental rejection is hypothesized to induce feelings of
unlovability and rejection in children.

Individuals who perceive themselves to be rejected are predicted to develop the
personality dispositions of a) hostility, aggression, or problems with the management of
both, b) dependence or defensive independence, c) impaired self-esteem, d) impaired
self-adequacy and self-esteem, e) emotional unresponsiveness or instability, and i) have a
negative worldview (Rohner & Britner, 2002; Rohner, 1991). A recent meta-analysis
using adults in the US. and internationally demonstrated that the weighted mean effect
size of the relation between the Adult Parental Acceptance-Rejection Questionnaire
(PARQ) and the Adult Personality Assessment Questionnaire (PAQ) is .46 using a
sample of 1,722 (Rohner & Britner, 2002). These results suggest that adults’ recollections
of childhood experiences with parental warmth (acceptance-rejection) explain
approximately 21% of the variance in their current psychological adjustment.

According to PARTheory, individuals who perceived themselves as rejected have
inherently different internal working models (mental representations of self and others)
resulting in a trajectory down a different developmental pathway, often resulting in
difﬁculty. Literature has demonstrated that when important affectional bonds are
threatened, distorted, or broken, psychological needs become unmet, thereby triggering
powerful emotions such as anxiety, insecurity and anger (Ward, Ramsay, Tumbull,

Benedettini, & Treasure, 2000).

38

A revisitation to the concept of “looking glass self” within symbolic
interactionism is helpful as it also arises as an assumption to PART. It states that
individuals tend to view themselves as they imagine signiﬁcant others view them.
Therefore, if parents constitute arguably the most signiﬁcant relationships in a child’s life
and thereby are perceived by the child to reject he or she, it is likely that these children
will deﬁne themselves as unworthy of love and have an overall negative self-evaluation
(Rohner, 1991). Consequently, a child perceiving themselves as rejected will engage in
behaviors such as seeking parental approval and attention, seeking physical contact and
being overly dependent in an effort to remedy the rejection (Rohner, 1991). The results of
prolonged perception of parental rejection is hypothesized through PART to facilitate less
tolerance for stress, emotional instability, and susceptibility for various
psychopathologies.

Also similar to the identity concept embedded within symbolic interactionism,
PART postulates that children with clearer senses of self and the ability to depersonalize
are better equipped to reject negative messages from parents (Rohner, 1991). Rohner
differentiates between copers and non-copers whereby copers are able to depersonalize
and deal more functionally with perceived rejection. However, PART does not clearly
deﬁne what constitutes coping. The construct of coping often engenders images of
strength and health. However, a variety of coping methods may be employed to deal with
painful situations or experiences that are less than desirable for one’s health such as
substance abuse or eating disorders. Therefore, the idea of coping being equated with
higher functioning in the proposed study whereby disordered eating may be called upon

as a consequence to parental rejection is intuitively inaccurate. Rather, a neutral

39

deﬁnition of coping is necessary to incorporate into the proposed study’s hypotheses
whereby coping may in fact be a reasonable result of parental rejection in the form of
maladaptive behaviors, such as disordered eating.

Both PART and symbolic interactionism advocate for a phenomenological
approach rather than a behaviorist one. With the assumption that human behavior is
affected to a greater extent by their perception or deﬁnition, both theories favor a stance
lending weight to subjective experience and interpretation rather than objectivity
(Rohner, 1986; LaRossa & Reitzes, 1993). Therefore PART emphasizes the child’s
experience of his or her parent’s warmth, hostility, or indifference.

Envoy (p. 34, 1981) links the presence of perceived parental rejection with the
incorporation of psychological devices to assist in coping in the following quote,

“No matter how obscure the explanations Of just how they had eventually come to

know they were rejected, one point seemed indisputable. A substantial number of

the rejected reported that, after they had ﬁnally recognized their rejection, for
years they unknowingly employed one or more psychological devices to protect
themselves from the unlovely realization that they had been rejected. These
psychological, rejection-concealing devices were “defense mechanisms,” which
involved no aspect whatever of conscious deception or pretending. Rather, they
were psychological tricks that these people had automatically and unconsciously
used in order to shield themselves from the exceedingly distressing experience of
rejection.”

Supporting PARTheory

To date, studies have implicated parental rejection in the etiology of drug and
alcohol abuse, schizophrenia, depression, conduct disorder, externalizing behaviors,
delinquency, and more recently, binge eating disorder (Khaleque & Rohner, 2002;
Rohner & Nielsen, 1978). Conceptually these symptoms may be viewed as preferred

ways Of coping when perceptions of rejection become overwhelming and lead to

intolerable emotions. In comparing children indicating perceived rejection with groups of

40

controls, rejected children have been shown to evaluate themselves more negatively and
be more dependent (Rohner, 1991).

The conceptualization of parental acceptance and rejection has an impressive
presence in the literature. To date, 1,500 studies exist documenting the assertion that
parental rejection is associated with varying forms of psychopathology, behavioral and
psychological problems, substance abuse, attachment disorders, academic difﬁculties,
and troubled personal relationships (Rohner, 2002). A number of these studies have
documented correlations between parental acceptance and pro-social behavior, positive
peer relationships, psychological well-bein g, and higher rates of life satisfaction and
happiness.

Steinhausen and Winkler Metzke (2001) conducted a study to target general risk
factors for mental disorders in an adolescent sample. The researchers found perceived
rejection by parents to be a risk factor for both males and females, co-varying with
problem behaviors as reported by the Youth Self Report (Steinhausen & Winkler Metze,
2001). Another study has looked at perceived relational support as related to adolescent
adjustment including psychological well-being, delinquency, and substance use (Scholte,
van Lieshout, & van Aken, 2001). In this study, factor analyses were used to understand
conﬁgurations of support, or lack there of, that were related to difﬁcult psychological
adjustment. The ﬁndings concluded that perceiving low support from parents was present
in each high-risk group of adolescents (Scholte, van Lieshout, & van Aken, 2001). One
aspect of the relational support construct used was warmth versus hostility involving
acceptance in the exchange between parent and child (Scholte, van Lieshout, & van

Aken, 2001). This deﬁnition is similar to how Rohner operationalizes parental rejection,

41

lying on a continuum of warmth with affection, support and caring on one end labeled
acceptance, and rejection on the Opposite pole involving a lack of this affection, support,
and caring (Rohner, 1986; 1991).

In an effort to display the universality Of parental acceptance-rejection theory,
PARTheory, Khaleque & Rohner (2002) conducted a meta-analysis utilizing ﬁfteen
studies for adult perceptions, both published and unpublished, looking at the relation
between adult scores on the Parental Acceptance-Rejection Questionnaire (PARQ) and
the Personality Assessment Questionnaire (PAQ). The mean weighted effect size of
maternal perception was r-=.46 and 1:45 for paternal perception indicating a moderate
effect. To cross check this ﬁnding once more, a fail safe N was computed to discover
how many additional studies would be needed to result in non-signiﬁcant results. The fail
safe N indicated that an additional 941 studies with non—signiﬁcant results would be
necessary to cancel the calculated effect size. Khaleque and Rohner (2002) draw the
conclusion based on the meta-analytic results that increases in perceived parental
rejection are positively associated with increases in psychological maladjustment.
Furthermore, in reviewing both transcultural and intracultural studies focusing on both
children and adult perceptions, a total of 43, the conclusion was drawn that no sample in
the world has yet been studied where PARQ and PAQ failed to produce the expected
positive correlations.

Several studies have attempted to better understand the relationship between the
perception of parental rejection and speciﬁc psychopathology. Emotional adjustment,
operationalized as an absence of anxiety and depression, in adolescence was studied

through this lens. This particular study examined the direct and indirect associations

42

between perceived parental acceptance and emotional adjustment, with the inclusion of a
third variable, perception of self-competence, hypothesized by the investigators to
mediate the direct relationship (McCauley, Ohannessian, Lerner, von Eye, & Lerner,
1996). The direct path model, including parental acceptance predicting emotional
adjustment, provided a good ﬁt. The indirect path model, including perceived self-
competence as a mediating variable, showed no statistically signiﬁcant paths. In regards
to the speciﬁc emotional adjustment outcome variable, the correlation between perceived
maternal acceptance and depression was found to be signiﬁcantly stronger for girls than
for boys, again implicating a need for split measures of parental acceptance-rejection
including maternal and paternal.

Another study looking at speciﬁc symptoms has been conducted by Campo and
Rohner (1992) as they investigated the connection between parental acceptance-rejection
to substance abuse. Findings included the substance abusing group felt signiﬁcantly more
rejected by both their fathers and mothers as compared to the group of non-substance
abusers (Campo & Rohner, 1992). Additionally, the study found that perceived paternal
and maternal acceptance—rejection was related signiﬁcantly to psychological adjustment
as reported using the Personality Assessment Questionnaire (Rohner, 1991). The results
of this study advocate for the importance of a developmental perspective focusing on the
long-term interaction effects of family dysfunction, perceived rejection, and
psychological maladjustment in the etiology of substance abuse. Using the analyzed data,
but recognizing that causation has not been achieved, the authors postulate that perceived

rejection may spark maladaptive personality and behavioral characteristics that may

43

further provoke rejection on the part of the parent, which may eventually lead to
substance abuse as a way of dealing with the emotional pain of perceived rejection.

Only one study to date has researched the effect of an individual’s perception of
parental acceptance or rejection on the occurrence of disordered eating. The results
indicated individuals with binge eating disorder perceived both their mothers and fathers
to be signiﬁcantly more rejecting than a group of controls (Dominy, Johnson, & Koch,
2000). However, a within group ﬁnding emerged that revealed obese women with binge
eating disorder reported more rejecting and less warm fathers than mothers. Using the
model proposed by Campo and Rohner (1992) in their investigation with the connection
between substance abuse and parental acceptance-rejection as a conceptual map, Dominy,
et al. (2000) postulated that binge eating would constitute a logical result of the parental
rejection process. Individuals may turn to one substance or another to decrease the
intense psychological pain resulting from perceived parental rejection. In the case of
binge eating disorder, food may be the substance of choice to numb this internal pain.

It is clear from the literature that although Rohner’s PARTheory has been found
useful in explaining numerous forms of maladjustment, it has yet to be empirically
researched in the populations of eating disorders including anorexia and bulimia. The gap
in literature pertaining to disordered eating individual’s perceptions of their parents levels
of acceptance or rejection is critical as it may be preventing the establishment of effective

treatment models.

CHAPTER 3
Methodology
Introduction

A good deal of research has addressed the connection between the perception of
parental acceptance- rejection and numerous emotional disturbances such as substance
abuse, depression, anxiety, conduct disorder, binge eating disorder, and externalizing
behaviors. The review of the literature suggests that individuals’ perceptions of
relationships with their parents play a signiﬁcant role in the development of eating
disorders. The proposed research focuses on perception of parental (maternal and
paternal) rejection and individual’s self-report of family functioning as variables that
distinguish between individuals exhibiting disordered eating attitudes and behaviors and
those without. Speciﬁcally, assessments were conducted to ﬁnd whether the perception of
parental acceptance-rejection and family functioning distinguish between individuals
with disordered eating tendencies and non-symptomatic individuals.

As evidenced by the literature, a gap, and therefore an opportunity, exists in
looking at the relationship between disordered eating symptomatology and the perception
of parental acceptance and rejection. Using research and theory as a guide, the eating
disorder literature is lacking in empirical attention paid to the interplay between Rohner’s
parental acceptance-rejection theory (Rohner, 1991) and the microsystemic processes
involving behaviors, thoughts, and attitudes found in individuals’ struggling with their
relationship with food and their body. At this time, a quantitative, empirically-based
study is necessary to ﬁnd if a relationship between the variables of family functioning,

parental rejection, and disordered eating exists.

45

In this chapter, the research questions and hypotheses will be revisited.
Conceptual and operational deﬁnitions of the study’s predictor and outcome variables
will be outlined, and variables will be introduced. The sampling procedure utilized in
recruitment of participants will be discussed. Instrumentation for the project will be
identiﬁed with each assessment being introduced complete with reliability and validity
data. Both data collection and data analysis will be outlined for the project, with Chapter
4 being reserved for a more in-depth discussion Of both, including results.

Research Questions & Hypotheses
As noted in Chapter 1, the following research questions are addressed in this study:
Research question 1:
Does parental acceptance-rejection differ between individuals with disordered eating
and those without?
Hypothesis 1:
Individuals presenting as disordered in their eating will perceive their parents as more
rejecting than individuals without disordered eating symptoms.
Research question 2:
Does family functioning differ between individuals with disordered eating and those
without?
Hymthesis 2:
Individuals presenting as disordered in their eating attitudes and behaviors will
perceive their families to be less healthy in terms of family functioning, than

individuals without disordered eating difﬁculties.

46

Research question 3:

Does the presence of parental rejection predict disordered eating symptomatology?
Hypothesis 3:

Individuals indicating more parental rejection will exhibit greater degrees of
disordered eating.

Research question 4:

Does the perception of unhealthy family functioning predict disordered eating
symptomatology?

Hymthesis 4:

Individuals indicating unhealthy family functioning will exhibit greater degrees of
disordered eating.

Research question 5:

Is maternal or paternal rejection a better predictor of disordered eating
symptomatology?

Hypothesis 5:

Although it is hypothesized that individuals reporting higher levels of disordered
eating will perceive both their mothers and father to be more rejecting than the non-
eating disordered group (hypothesis 1), it is further hypothesized that individuals
reporting disordered eating symptomatology will perceive their fathers to be more

rejecting than their mothers.

47

Conceptual Model

   

 
    
   
   
 
       
   
 
   
   

ERCEPTIO
OF PATERNAL
ACCEPTANCE-

REJECTION

 

 

PERCEIVED
UNHEALTHY DISORDERED
FAMILY EATING

FUNCTIONING

 

 

 

 

PERCEPTION OF
MATERNAL

Figure 2 Conceptual Model
Conceptual and Operational Deﬁnitions
0 Parental rejection

Conceptual: Absence of parental warmth and affection; rejection lies on one end
of a warmth continuum with acceptance at the other. Parental rejection is not a speciﬁc
set of actions or behaviors displayed by parents, but rather the perception by the child that
he or she is unloved.

Operational: For the purpose of this study, parental rejection was measured by the
Parental Acceptance-Rejection Questionnaire (PARQ). Scores for the survey range from
60-240 for each parent (Rohner, 1991). A score of 240 represents the maximum
perceived rejection score. Individuals scoring higher than 150 were interpreted as

perceiving more rejection than acceptance from the rated parent. Scores between 140 and

48

149 reveal that respondents experience serious rejection, but not necessarily more overall
rejection than acceptance (Khaleque & Rohner, 2002).
0 Parental acceptance

Conceptual: Presence of parental warmth and affection, leading to a child’s
perception of self as loved and accepted by his or her parent(s); acceptance lies on the far
end of the warmth continuum, representing the opposite pole of rejection. Parental
acceptance may be evidenced through physical and/or verbal affection as perceived by
the child as indicating love and caring from the parent(s).

Operational: For the purpose of this study, parental acceptance was measured
using the Parental Acceptance-Rejection Questionnaire. A score of 60 represents the
maximum perceived acceptance score. Individuals scoring less than 150 were interpreted
as perceiving more overall acceptance than rejection from the rated parent (Rohner,
1991).

0 Family functioning

Conceptual: Family functioning is a multidimensional construct deﬁned as the
accomplishment of essential functions and tasks by family members in a system
(Grotevant & Carlson, 1989).

Operational: For the purpose of this study, family functioning was assessed using
the Family Assessment Device (Epstein, Bishop, & Levin, 1978; Epstein, Baldwin, &
Bishop, 1983; Kabacoff, Miller, Bishop, Epstein & Keitner, 1990). Seven dimensions are
measured using the FAD and include: problem solving, communication, roles, affective
responsiveness, affective involvement, behavior control, and general functioning. The

composite score of the FAD’s seven scales was used for the proposed study. A higher

49

score indicates more unhealthy perceptions of interaction within the family, noted as
family dysfunction for the purposes of this study (Waller, 1994).
0 Disordered eating and non-disordered eating

Conceptual: Eating attitudes and behaviors are conceptualized as lying on a
continuum with asymptomatic eating on one end (non-disordered eating) and
symptomatic on the other (disordered eating). Disordered eating is used in this study to
describe an individual who engages in behaviors commonly associated with eating
disorders such as food restriction, purging, preoccupation with food and appearance, and
excessive exercising.

Operational: For the purpose of this study, presence of disordered eating
was measured through the Eating Attitudes Test-26 (Garner, Olmsted, Bohr, & Garﬁnkel,
1982). The EAT-26 was designed to measure both behavioral and attitudinal symptoms
associated with the eating disorders of anorexia nervosa and bulimia nervosa. Individuals
scoring 20 or higher are considered to be disordered in their eating attitudes and
behaviors, while individuals scoring less than 20 are considered non-disordered.

Sampling Procedure

The sample for this study consists of female undergraduate college students
attending a large Mid-western university. Research has shown that prevalence rates of
eating disorders are higher in college students than in other samples (Vohs, Heatherton,
& Herrin, 2001; Gutzwiller-Jurman, 1999; Fairbum & Beglin, 1990), with this age group
exhibiting higher rates of dieting, body dissatisfaction, and disturbed eating (Heatherton,
Mahamedi, Striepe, Field, & Keel, 1997). College women between the ages of 18 and 22

have higher rates Of bulimia than those females younger, not in college, or over twenty-

50

one years (Gutzwiller—Jurman, 1999). A purposive, non—probability sampling procedure
was used in an effort to target a non—clinical population more likely to exhibit disordered
eating behaviors than the general population. The incidence of eating disorders on
university campuses is strikingly high with studies showing prevalence rates for females
ranging from 4 to 9% (Hesse-Biber, Marino, & Watts-Roy, 1999; Pope, Hudson,
Yurglen-Todd, Hudson, 1984; Schwitzer, Bergholz, Dore, & Salimi, 1998). For women,
weight and body dissatisfaction were found to be signiﬁcantly higher during college as
compared to post-college reports (Vohs, Heatherton, & Herrin, 2001). One study found
one-quarter of female undergraduates reported their eating to be out-of-control
(Koszewski, Newell, & Higgins, 1990).

Inclusionary criteria for the study required individuals to be enrolled as
undergraduate students at Michigan State University, female, and between the ages of 18
and 25 years. In an effort to increase homogeneity in the sample, students not meeting
these criteria were not included in the analysis.

A non-clinical sample is desirable for the present study as it is exploratory in
nature, looking to better understand the relation between varying degrees of parental
rejection and family functioning with eating behaviors and attitudes. A non-clinical
sample provides a range of disordered eating behaviors with which to explore two
groups: disordered eating and non-disordered eating. Additionally, a larger N was
possible with a non-clinical versus a clinical sample, allowing for greater ﬂexibility with
statistical analyses.

Based on a study looking at the prevalence of eating attitudes and behaviors using

the Eating Attitudes Test-26 in an undergraduate college sample, 20% of the females

51

scored above the EAT-26 cut-Off score, indicating symptomatology associated with
anorexia nervosa including dieting, bulimia, food preoccupation, and oral control
(Nelson, Hughes, Katz, & Searight, 1999).
Variables

Three primary independent variables in this study include perceived level of
maternal acceptance-rejection, perceived level of paternal acceptance-rejection, and
perception of family functioning. Both parental rejection-acceptance and family
functioning variables are continuous scales allowing for ease in incorporation into a
multiple regression model. Beyond the primary independent variables to be used in an
abbreviated regression model, each of the composite variables, family functioning and
parental acceptance-rejection, also include sub-scale variables. The inclusion of these
sub-scales as replacement of the composite primary variables resulted in an expanded
regression model, lending to the speciﬁcity of the model. For instance, the parental
acceptance-rejection composite variable (both maternal and paternal) is made up of four
sub-scale variables including warmth/affection, aggression/hostility, neglect/indifference,
and undifferentiated rejection. Additionally, the family functioning variable as
operationalized by the Family Assessment Device (Epstein, Bishop, & Levin, 1978)
includes seven sub-scales, with general functioning being used as the primary
independent variable for the abbreviated model discussed above. All seven sub-scale
variables were incorporated into the expanded model to specify variable relationships.
The six additional sub-scales from the Family Assessment Device (in addition to the
general functioning scale) include problem solving, communication, roles, affective

involvement, affective responsiveness, and behavior control. To summarize, an

52

abbreviated model included three primary independent variables and an expanded model
was run with 15 sub-scale independent variables as a means of increasing model
speciﬁcity.

The outcome, or dependent variable, was level of disordered eating as measured
by the Eating Attitudes Test-26 (EAT-26). The EAT-26 incorporates a continuum,
measuring a full range of eating behaviors and attitudes ranging from asymptomatic (non-
disordered) to symptomatic (disordered). Researchers have viewed eating disorders as
including anorexia nervosa and bulimia nervosa as an extreme end Of a continuum based
on an individual’s relationship with food (Hesse-Biber, Marino, Watts-Roy, 1999;
Striegel-Moore, 1992). A continuum for categorizing disordered eating has been selected
for this study based on literature that has found partial syndromes to be two to ﬁve times
more common than full diagnosis in adolescent females, ranging from 4% - 16% of the
general population (Mussell, Binford & Fulkerson, 2000). Additionally, between 60-80%
of college women are reported to engage in regular bingeing and other disturbed
behaviors that do not qualify for full clinical diagnoses (Hesse-Biber, Marino, Watts-Roy,
1999). The importance for attention paid to partial syndromes or disordered eating has
been supported by empirical evidence advocating for continuity between all levels than
distinctly considering one relevant and the other not (Stice, Killen, Hayward, & Taylor,
1998).

In addition to offering the ﬂexibility of a continuous variable of eating attitudes and
behaviors, the EAT-26 also provides a cut-Off whereby two groups may be created

including eating disordered and non-eating disordered. This dichotomous variable was

53

useful for data analysis using t-tests, whereby the continuous nature of the variable will
be incorporated into the multiple regression model.
Instrumentation

Demographic Questionnaire: An investigator-derived demographic questionnaire
was used to gather data on participants’ gender, age, family income level, parents’
relationship status, and ethnicity (see Appendix B). Demographic information gathered
from the questionnaire will be used for descriptive purposes of the sample.

Eating Attitudes Test-26: The outcome variable examined within the proposed
study consists of eating attitudes and behaviors. Presence of disordered eating was
measured through the Eating Attitudes Test-26 (Garner, Olmsted, Bohr, & Garﬁnkel,
1982). The EAT-26 does not yield a speciﬁc eating disorder diagnosis, but it has been
deemed useful in detecting eating disturbances in non-clinical samples, making it
particularly useful to the proposed study’s aims. The EAT—26 differentiates individuals
with anorexia and bulimia nervosa from controls; however, the symptomatic individuals
score in the same range (Netemeyer & Williamson, 2001).

The EAT-26 was derived from the original EAT—40 using factor analysis going
from seven scales and 40 questions to three scales and 26 questions. Both measures were
found to be correlated at #098, indicating that the 14 deleted questions were redundant
and not necessary for maintaining the predictive power Of the instrument. Cronbach’s
alpha, indicating the reliability and internal consistency of the EAT-26 has been shown to
be 0.90 for the eating disordered group and 0.83 for the non-disordered eating group

(Garner, et al., 1982).

54

Garner and Garﬁnkel (1979) constructed the BAT to reﬂect a range of anorexic
behaviors and attitudes. Three factors, or subscales, make—up the EAT-26 and include
dieting, bulimia and food preoccupation, and oral control (Garner, Olmsted, Bohr, &
Garﬁnkel, 1982). In a sample of 160 female eating disorder patients with half being
restrictors and the other half being classiﬁed as bulimic subtypes, no difference was
found in the two groups’ composite EAT-26 scores (Garner, Olmsted, Bohr, & Garﬁnkel,
1982). Therefore, individuals scoring above the speciﬁed cut-off for the instrument can
be classiﬁed as disordered eating. The composite EAT-26 score will be used in the study
rather than the three subscale scores.

The EAT-26 includes items that are rated on a 6-point Likert scale ranging from 1
indicating always and 6 indicating never. The EAT-26 is designed to measure both
behavioral and attitudinal symptoms associated with the eating disorders of anorexia
nervosa and bulimia nervosa. The authors report internal consistency (coefﬁcient alpha)
of .90 and a discriminant analysis resulting in 84% correct classiﬁcation of known
groups. Again, respondents are instructed to answer using a Likert scale with responses
including always, usually, often, sometimes, rarely, and never. Sample items from the

EAT-26 include:

0 Am terriﬁed about being overweight.

0 Find myself preoccupied with food.

0 Have gone on eating binges where I feel that I may not be able to stop.
0 Have the impulse to vomit after meals.

0 Feel extremely guilty after eating.
A cut-off score of 20 is suggested for determining symptomatic individuals versus

non-symptomatic individuals (scores >20). The measure yields one composite score,

55

 

E.\1!I.1J r

 

classifying an individual on a continuum of eating with scores above 20 classiﬁed as
disordered. Therefore, a score Of 19 would not be considered disordered using the EAT-
26 developer’s score Operationalization and interpretation. Scoring for the six possible
responses is as follows: always = 3; usually = 2; often = 1; sometimes, rarely and never =
0. One item (25) is reverse scored.

The cut-off score of 20 for the EAT-26 was derived from the cut-off established
for the EAT—40, which was determined to be 30. A cut-off of thirty was used to eliminate
false negatives using data that suggested 7% of the normals scored above the lowest
active symptomatology score (Garner & Garﬁnkel, 1979). When the EAT-26 was created
as a shorter version of the original EAT-40, the cut-off score was shifted to 20. Garner, et
a1. (1982) found this cut-off to correctly classify a similar proportions of disordered
eating and normals when compared to the cut—Off of 30 for the EAT-40. The EAT-26 cut-
off score has been shown to identify approximately 15% of college students (Garner, et
al., 1982). Discriminant function analysis was used to ﬁnd the percentage of cases
correctly classiﬁed to be 84.9% for the EAT—40 and 83.6% for the EAT-26 (Garner, et
al., 1982).

This study conceptualized eating on a continuum from non-disordered to
disordered, allowing for the ﬂexibility of focusing on disordered eating as a larger
category beyond the discrete diagnostic categories of anorexia nervosa and bulimia
nervosa. Striegel-Moore, Silberstein, Frensch, and Rodin (1989) acknowledge that even
at severity levels that fall considerably below diagnostic criteria for eating disorders,

maladaptive dieting patterns and eating warrant attention.

56

Additional analyses are possible through utilization of the EAT-26’s three sub-
scales which include dieting, bulimia and food preoccupation, and oral control. These
three subscales may be used as more speciﬁc dependent variables, evidencing behaviors
correlating with eating disorder diagnoses. These analyses may be desirable for future
research in this area as other researchers have often found females struggling with
bulimic symptoms (bingeing and purging) to report greater family dysfunction including
more distant, conﬂictual, and non-supportive (Humphrey, 1989; Johnson & Flach, 1985).
Another study found patients with anorexia (purging subtype) were more likely to
perceive family functioning as impaired than were either controls or restricting anorexic
patients (Casper & Troiani, 2001). Considering the exploratory nature of the proposed
study and the use of the EAT-26, which does not result in a clinical diagnosis, differential
behaviors corresponding to speciﬁc diagnoses will not be incorporated into the study as

dependent variables.

Parental Acceptance and Rejection Questionnaire (Adult — PARQ): Parental

 

acceptance-rejection was measured by the adult version of the Parental Acceptance and
Rejection Questionnaire (PARQ). The adult-version Parental Acceptance/Rejection
Questionnaire (PARQ-F, father version; PARQ-M, mother version), a self-report measure
developed by Rohner (1986, 1991), was designed to assess an individual's perception of
acceptance-rejection by his or her parents during childhood, speciﬁcally during the ages
of 7 through 12 years of age. Participants were provided instructions to complete the
questionnaire for their biological mothers and fathers. Participants who were unable to
follow these instructions due to no contact with their mother or father growing up did not

complete the questionnaire. There was one participant who did not complete the

57

maternal acceptance-rejection questionnaire and an additional nine did not complete the
paternal acceptance-rejection questionnaire. It is a 60-item, Likert-type inventory with
possible responses including almost always true, sometimes true, rarely true, and almost
never true. Sample items from the Parental Acceptance-Rejection Questionnaire include,

“My Mother/Father”:

o Takes an active interest in me...

0 Ignores me when I ask for help. . .

0 Says nice things to me when I deserve them...
0 Likes to spend time with me...

o Feels other children are better than I am no matter what I do...

Scores on the four subscales--parental warmth, hostility, neglect, and
undifferentiated rejection--are combined to determine a composite score, which can range
from 60 to 240 (midpoint = 150). A score of 60 represents the maximum perceived
acceptance score and a score of 240 represents the maximum perceived rejection score.
This composite score was utilized in the proposed study. Individuals scoring higher than
150 are interpreted as perceiving more rejection than acceptance from the rated parent.
Scores between 140 and 149 reveal that respondents experience serious rejection, but not
necessarily more overall rejection than acceptance (Khaleque & Rohner, 2002). In
contrast, a score between 60 and 120 reveals a perception of substantial parental
acceptance and love (Khaleque & Rohner, 2002). Participants will be asked to complete
the instrument for both their mother and father.

Reliability studies (Rohner, 1991) have yielded Cronbach's alpha coefﬁcients for
the subscales ranging from .86 to .95. Mean test-retest reliability is .62 across time

periods ranging from 3 weeks to 7 years (Khaleque & Rohner, 2002). To test the

58

concurrent validity of the PARQ, a modiﬁed version of the instrument was produced by
inserting items from two already validated instruments including the Child’s Report of
Parent Behavior Inventory and Bronfenbrenner’s Parental Behavior Questionnaire. These
two measures were used as external, criterion measures. Results indicate that all four
scales of the PARQ are signiﬁcantly related to their respective validation scales at a level
of p<.001. Speciﬁcally, r values for the PARQ’s concurrent and criterion values are as
follows: warmth/affection scale = .90; aggression/hostility scale = .43,
neglect/indifference = .86, and rejection/undifferentiated = .81 (Rohner, 1991).

family Assessment Device (FAD): The remaining predictor variable of
perception Of family functioning was assessed through the Family Assessment Device
(FAD). This instrument is derived from the McMaster Model of Family Functioning,
specifying the farme dimensions Of problem solving, communication, roles, affective
responsiveness, affective involvement and behavior control (Epstein, Bishop, & Levin,
1978; Epstein, Baldwin, & Bishop, 1983; Kabacoff, Miller, Bishop, Epstein & Keitner,
1990).

The model’s various six dimensions, excluding general functioning, are
operationalized as follows. Problem solving refers to the family’s ability to resolve
problems that maintains effective family functioning. The exchange of information
among family members is measured by the communication sub-scale. The roles
dimension includes consideration of whether tasks are clearly and equitably assigned to
family members. Affective responsiveness assesses the ability to respond to an event or
situation with the appropriate quality and quantity of feelings (Epstein, Bishop, Ryan,

Miller, and Keitner, 1993). An additional affective dimension, aﬂective involvement,

59

measures the extent to which the family exhibits interest in and value activities of its’
members. Finally, the behavior control subscale assesses how the family creates and
maintains standards for it’s members’ behavior.

The seven dimensions identiﬁed above make—up a 60-item inventory used to
evaluate an individual’s perception of their family, which takes approximately twenty
minutes to complete (Miller, Bishop, Epstein, & Keitner, 1985; Epstein, Baldwin, &
Bishop, 1983). The response format uses a four point Likert-scale ranging from strongly
agree to strongly disagree (Epstein, Baldwin, & Bishop, 1983). A higher score indicates
more unhealthy perceptions of interaction within the family (Waller, 1994). Sample items
from the Family Assessment Device include:

0 If someone is in trouble, the others become too involved.

0 We resolve most everyday problems around the house.

0 People come right out and say things instead of hinting at them.
0 We avoid discussing our fears and concerns.

0 Each of us has particular duties and responsibilities.

The FAD has been shown to discriminate between clinical and non-clinical
families using a cut-off score of 2 on the general functioning scale to identify families
with signiﬁcant adjustment difﬁculties (Kabacoff, et al., 1990). Additionally, using
clinician’s ratings as criteria, cut-offs have showed acceptable rates of sensitivity (57-
83%) and speciﬁcity (64-79%) (Miller, Epstein, Bishop, & Keitner, 1985). The following
FAD cut-off scores have been found to discriminate signiﬁcantly between families with a
psychiatric illness and non-clinical families (Miller et al., 1985): problem solving 2.2,

communication 2.2, roles, 2.3, affective responsiveness 2.2, affectiveness involvement

60

2.1, behavior control 1.9, and general functioning 2.0. Again, a higher score indicates
more unhealthy perceptions of interaction within the family.

The Family Assessment Device has been demonstrated to have adequate internal
consistency and test-retest reliability (Miller et al., 1985). Internal consistency across the
seven scales using Cronbach alpha scores range from .72 to .92 (Grotevant & Carlson,
1989). Another study found alpha scores to range between .57-.83 for the non-clinical
families across the seven scales and 69-86 for psychiatric and medical families in their
sample (Kabacoff, Miller, Bishop, Epstein, & Keitner, 1990). The roles sub-scale is the
only dimension showing marginal reliability (.69 alpha for clinical samples and .57 for
non-clinical samples).

In terms of concurrent validity, a discriminant analysis of individual measure
scores predicted 67% of the non-clinical group and 64% of the clinical group with a
signiﬁcance level of p<.001 (Epstein, Baldwin, & Bishop, 1983). Finally, coefﬁcients of
factor invariance ranged from 95-99 across samples and factors. These indicators
provide strong evidence for a stable and consistent factor structure across all three
groups: psychiatric, medical and non-clinical (Kabacoff et al., 1990).

TABLE 1: Abbreviated Model - Variables and corresponding Instruments

 

 

Variables Level Type Instruments
Eating Attitudes and Continuous
Behaviors/Level Of Disordered . Dependent Eating Attitudes Test
Eating Categorical

 

Parental Acceptance-Rejection
(matemal & paternal)

Parental Acceptance and

Continuous Inde ndent . . . .
pe Rejection Questronnarre

 

Family Functioning Continuous Independent Family Assessment Device

 

 

 

 

I_

 

 

61

TABLE 2: Expanded Model - Variables and corresponding Instruments

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Variables Level Type Instruments
Eating Attitudes and Continuous
Behaviors/Level of Disordered C . al Dependent Eating Attitudes Test
Eatin ategorrc
g

Warmth/Affection Continuous Independent Parental Acceptance a?"

Rejection Questronnarre

. . . . Parental Acceptance and

Aggressron/Hostrlrty Continuous Independent Rejection Questionnaire

. . Parental Acceptance and

Neglect/Indrfference Continuous Independent Rejection Questionnaire

Undifferentiated Rejection Continuous Independent Parental Acceptance “."d

Rejectron Questionnaire
General Functioning Continuous Independent Family Assessment Device
Communication Continuous Independent Family Assessment Device
Problem Solving Continuous Independent Family Assessment Device
Roles Continuous Independent Family Assessment Device
Affective Involvement Continuous Independent Family Assessment Device
Affective Responsiveness Continuous Independent Family Assessment Device
Behavior Control Continuous Independent Family Assessment Device

 

 

 

 

Data Collection

Following the attainment of permission by the University’s Human Subjects

Institutional Review Board to conduct the research project, the principal investigator

established on-line access to the study’s questionnaires through the Department of

Psychology’s Human Subjects Pool. Students enrolled in various undergraduate

 

psychology classes can log onto this site and complete any number of research projects
activated for course credit. Additionally two instructors in the Department of Family and
Child Ecology were approached about recruitment through their live courses. One
instructor agreed to participate in the study, and further agreed to offer extra credit to any
of her students who completed the study’s questionnaires. The investigator Offered no
additional incentives to participants.

Two recruitment methods were chosen in order to obtain a more representative
female student sample. All students who agreed to participate in the study were provided
a list of professionals specializing in the treatment of disordered eating. This list included
nutritionists, psychotherapists, and psychiatrists easily accessible to students for both
geographical and ﬁnancial reasons. The referral list was provided to all study participants
rather than only “high risk” respondents based on EAT-26 scores, because Of
conﬁdentiality measures employed. Students only provided their name for the purpose of
informed consent. All other measures contain numerical coding, making contacting the
participants impossible.

Participants were asked to complete ﬁve questionnaires including: 1)
demographic information, 2) patemal-version parental acceptance rejection
questionnaire, 3) matemal-version parental acceptance rejection questionnaire, 4) family
assessment device, and 5) eating attitudes test. Time estimated to complete the entire
battery of assessments was 1 hour with the majority completing the assessments in no

longer than 40 minutes.

63

Data Analysis

Two primary statistical procedures were used in analyzing the data including t-

tests and multiple regression analysis. The SPSS program was used for data organization

and analysis. SPSS is a commonly accepted statistical program within the social sciences.

 

 

 

 

 

 

 

symptomatology?

 

maternal rejection.

 

TABLE 3:
Research Question with corresponding Hypothesis and Data Analysis
. . Data Analysis
Research Question Hypothesrs Technique

1. Does parental acceptance— Individuals presenting as disordered
rejection differ between eating will perceive their parents as T- tes t
individuals with disordered more rejecting than individuals
eating and those without? without disordered eating symptoms.

2. Does family functioning Indlv1duals presenting SS disordered

. . . . eating wrll perceive their families to
differ between indIVIduals . .
. . . be less healthy in terms of fanuly T-test
wrth disordered eating and . . . . . .
. functioning, than Indiv1duals Without
those Without? . .
disordered eating symptoms.

3. Does the presence of parental Individuals indicating more parental Multiple
rejection predict disordered rejection will exhibit greater degrees Regression
eating symptomatology? of disordered eating. Analysis

4' Does the ”“89““ 0f. . Individuals indicating unhealthy Multiple
unhealthy family functioning . . . . . . .

. . . family functioning wrll exhibit Regressron
predict disordered eating . . .
greater degrees of disordered eating. Analysrs
symptomatology?

5' Ismaternal or paternal Paternal rejection is a better Multiple
rejection a better predictor of . . . .

. . predictor of disordered eating Regressron
disordered eating .
symptomatology as compared to AnalySIS

 

T-tests were used to address research questions 1 and 2. Difference in means

between the disordered eating and non-disordered groups based on the independent

variables were attained. It was hypothesized that the disordered eating group, established

by using the cut-off of 20 on the EAT-26, would include higher parental rejection means

and higher unhealthy family functioning means than the non-disordered eating group.

 

Multiple regression analysis were used to compute the amount of variance
predicted by the independent variables on the dependent variable, level of disordered
eating. Regression analysis was useful in determining a model offering the best ﬁt with
the identiﬁed independent variables. Research questions 3, 4, and 5 were addressed using
multiple regression analysis.

The dependent variable in the model is the level of disordered eating measured by
the Eating Attitudes Test. Throughout the project, this variable is referred
interchangeably as disordered eating and eating attitudes and behaviors. The EAT allows
for both continuous and categorical variable scales with the measures creators identifying
the presence of disordered eating with scores above 20. For the purpose of the regression
model however, the dependent variable will be used as having a continuous scale to
measure levels of disordered eating with scores ranging from 0 — 79, with lower scores
indicating normal eating attitudes or an absence of disordered eating and higher scores
indicating more disordered eating behaviors and attitudes.

The ﬁrst and second independent variables in the model are the parental
acceptance-rejection variables as assessed using the paternal and maternal-versions Of the
Parental Acceptance/Rejection Questionnaire. Because a maternal and paternal version of
this questionnaire was used, both variables were entered into the model in order to
discover if one accounts for more variance than the other in explaining the dependent
variable, disordered eating attitudes and behaviors.

The third independent variable in the model is family functioning as perceived by
the participant and measured by the general functioning scale Of the Family Assessment

Device.

65

CHAPTER 4
Results
Introduction

In this chapter, sampling and data collection will be discussed ﬁrst as a foundation
for the study’s methodology. The principal investigator speciﬁes the data cleaning
process and identiﬁes the resulting sample size and characteristics of the sample
according to demographics. Frequency distributions are provided for the sample based on
each of the independent and dependent variables. Additionally, correlations and internal
consistency scores are reported for the independent and dependent variables before
moving on to the analysis to address the study’s hypotheses. T-test results will be
reported in response to hypotheses l and 2, and regression analyses were incorporated to
address hypotheses 3, 4, and 5. Four total regression models will be reported in an effort
to capture the complexity of the interpretation and speciﬁcation of the models. Of these
four there will be two abbreviated models (1 enter method and l step-wise), one
expanded model, and one follow-up to the expanded model to check for any effects of
collinearity.

Data Collection and Sample Recruitment

Between March of 2004 and May of 2004, the project’s questionnaires were made
available for university students enrolled in a variety of undergraduate psychology
classes to access and complete via computer participation. The students were offered
course credit for selecting and participating in the research study as established through
the psychology department at Michigan State University. The total number of participants

enrolling in the study via computer-based questionnaire completion was 1001. A second

66

method of subject recruitment included targeting students enrolled in an introductory
Family and Child Ecology course. The investigator attended a class session, was invited
to explain the study, and invited students to participate. The course instructor offered
extra credit to students willing to participate. A total of 168 participants were
accumulated using this recruitment method bringing the total sample size to 1169.

Three stages of data cleaning were used based on inclusionary criteria and
incomplete data to omit 335 of the 1169 original subjects. Nine were omitted because
they exceeded the study’s predetermined age criteria, falling into a category of older than
25. An additional 320 male subjects were eliminated from analysis because they did not
meet the inclusionary criteria of being female. And six subjects were omitted from
analysis due to incomplete Eating Attitudes Tests, viewed by the researcher as necessary
data because of its measurement of the study’s dependent variable. A total of sample size
of 834 resulted from the data cleaning process. This sample size is relatively large in
comparing it to other research studies using non-clinical, college samples which have
yielded subjects ranging from 101-249 in number (Thomson Ross & Gill, 2002;
Twamley & Davis, 1999).

The total resulting sample was expected to be homogeneous because of
inclusionary criteria established by the investigator (age, enrollment in university class,
and sex) as well as the sampling parameters including a college campus. However,
because two recruitment techniques were utilized to acquire the sample, analyses were
conducted to ensure no differences exist. The participants were each given a code of
“computer” or “classroom” to indicate how they enrolled in the study. A t-test was

conducted to assess for potential differences between the two samples based on the

67

study’s dependent variable, eating attitudes and behaviors as measured by the Eating
Attitudes Test-26.

A total of 692 individuals enrolled in the study via computer recruitment and 142
enrolled through classroom recruitment. Results from Levene’s test for equality of
variances indicate that the samples have homogeneous variances (F=3.501, p=.062). The
mean score on the EAT for group one (computer recruited) was 12.1156, whereas the
mean score for group two (classroom recruited) was 10.6901. T-test analysis found this
mean difference to be non-signiﬁcant (t=1.307, p=. 192). Based on this analysis the two
recruitment samples were combined into one homogeneous sample for the study’s data
analysis.

Demographic sample characteristics

Frequency analysis was used to generate demographic data for the sample Of 834
young women. The sample is predominately Caucasian (80.3%), between the ages of 18-
19 (57%) and currently of Freshman status in college (50.1%). Ethnicity data indicate the
second largest group represented is African-American, accounting for 7.9%, followed by
Asian-Paciﬁc Islander (4.9%) and Hispanic American (2.5%). A sizable portion of the
sample (3.7%) either described themselves as “other” regarding ethnicity or were placed
in this category by the investigator if two or more backgrounds were identiﬁed.
Therefore, many of the subjects within the “other” ethnic group are bi- or multi-racial.
According to student census data reported from Michigan State University’s Ofﬁce of the
Registrar for Spring semester 2004, 75.6% of the enrolled study body were classiﬁed as
Caucasian, 8.1% African American, 4.9% Asian American, 2.7% Hispanic and Mexican

American, and .6% Native American. In comparing the study’s sample to the overall

68

university sample, the ethnic group representations are nearly identical and the study
sample can therefore be considered representative of the larger campus population based
on this variable.

Nearly half (49.3%) of the sample came from families earning $75,000 or more
per year. Family structural data indicate that the majority (63.3%) of subjects’ biological
parents are married and living together with the next largest portion identifying their
parents as divorced (19.1%). Seventy two percent of the sample reports having one or

two siblings, 22 percent indicate 3-6 siblings and 6 percent report being an only child.

69

TABLE 4: Summary Demographic Characteristics of Sample

 

 

 

 

 

Demographics Number Percentage_
_A_ge (years)
18-19 476 57
20-21 291 35
22-23 61 7.3
24-25 6 .7

 

 

Year in colleg

 

 

 

 

 

 

 

 

 

 

 

 

Freshman 41 8 50. 1
Sophomore 194 23.3
Junior 141 16.9
Senior 81 9.7
Ethnicity

Caucasian 670 80.3
African-American 66 7.9
Asian American/Paciﬁc Islander 41 4.9
Hispanic American 21 2.5
Native American 5 .6
Other 3 1 3.7

 

 

Family income

 

 

 

 

 

 

$15,000 or less 39 4.7
$15,001-$35,000 74 8.9
$35,001 -$50,000 127 15.2
$50,001-$75,000 179 21.5
$75,001 or above 411 49.3
Unknown 4 .5

 

 

Biological parents’ relationship

 

 

 

 

 

 

Married 528 63.3
Not married but living together 9 1.1
Separated 12 1.4
Divorced 159 19.1
Widowed 29 3.5
Never married 97 11.6

 

 

 

Variable distributions across the sample
Frequency data regarding the Eating Attitudes Test, used to measure the

dependent variable in the study reveals 79 percent (n=659) scoring below 20, indicative

70

of non-disordered eating and 21 percent (n=175) scoring at or greater than 20, identifying
individuals at high risk for development of an eating disorder. Possible EAT composite
scores range from 0 to 79; the minimum score within the study’s sample was 0 and the
maximum score was 69. The mean EAT composite score was 11.87 (SD: 11.84) and the
median was seven. The group scoring 20 or greater in this study will be referred to as
disordered eating.

In an early standardization study completed by the EAT-26 developers, (Garner,
Olmsted, Bohr, & Garﬁnkel, 1982) the EAT-26 cut-Off score of 20 was shown to identify
approximately 15 percent of college students. However, a more recent study identiﬁes the
proportion of disordered eating subjects more closely to this study, approximately 20
percent (Nelson, Hughes, Katz, & Searight, 1999). This increase in subjects scoring as
higher risk for disordered eating may be a reﬂection of the increasing trend of eating

disorders and extreme dieting behaviors on college campuses across the United States.

 

300

200‘

 

 

 

100
>~

3 Std. Dev: ”.34
a)

a. Mean: 11.9

d)

u‘: o N=834.00

0.0 10.0 20.0 30.0 40.0 50.0 60.0 70.0
5.0 15.0 25.0 35.0 45.0 55.0 65.0

EATCOMP
Figure 3: Distribution of EAT-26 Composite Scores

71

Internal consistency of the Eating Attitudes Test was estimated using Cronbach’s
alpha whereby or = .9057. A standardization study by Garner and colleagues (1982) found
the EAT-26 to demonstrate an alpha of 0.90. Analysis of speciﬁc questions included on
the EAT-26 is included in the Clinical Implications section of Chapter 5.

The independent variable of family functioning was measured using the Family
Assessment Device (FAD). The FAD is divided into “healthy” and “unhealthy”
functioning items. “Unhealthy” functioning items were reverse scored to deliver
consistency in interpretation. Although the FAD is made up of 7 sub-scales, only the
General Functioning scale would be used for this study. The FAD does not recommend
using a total composite score using all seven subscales. A composite score was derived
by summing the items within the General Functioning scale and then dividing by the total
number of items in the scale, creating a mean score. This mean score serves as the scale
composite for each participant.

A possible range of 1.00 to 4.00 exists for the General Functioning composite
scores with 1.00 indicating greater healthy functioning as perceived by the respondent
and 4.00 indicating more unhealthy functioning (dysfunction) as perceived by the
respondent. The “general functioning” sub-scale is the only one with a speciﬁed cut-off
with previous research indicating a score of 2.00 or above identiﬁes problematic family
functioning; the higher the score, the more problematic the participant perceives her
family’s overall, or general functioning.

The General Functioning variable (GFAD) was recoded using the original
continuous scale into a dichotomous scale to identify the ratio of problematic to non-

problematic families perceived in the study. Non-problematic families were coded to

72

include participants who scored less than 2.00 on the General Functioning sub-scale of
the FAD and problematic families were coded to include participants scoring 2.00 or
greater on the GFAD. Nearly 67 percent (n=496) of the study sample perceived their
families to be healthy or non-problematic with approximately 33 percent (n=338)
perceiving their families to be unhealthy or problematic according to the General
Functioning scale of the FAD (see Table 5 for frequency distribution of GFAD).

TABLE 5: Frequency Data for the FAD General Functioning Scale

 

 

 

Frequency Percent
General Functioning
(healthy) 496 59.47
<2.00
General Functioning (unhealthy) 338 40. 53
2.00 or <
Total 834 100

 

 

The mean score on the GFAD was 1.859 (SD .580). The minimum score was 1.00
and the maximum score was 3.92. As evidenced by the histogram below, the distribution
of scores was more normal than would have been predicted or expected. The mode for
the GFAD was 2.00, indicating problematic family functioning. The ability of the FAD to
clearly differentiate problematic from non-problematic families may be part of the reason
for inconclusive results in the literature regarding its relationship with disordered eating
symptomatology. Cronbach’s alpha for the GFAD was demonstrated to be high with an

alpha of .9226.

73

 

   

 

>.

2 Sid. Dev = .58
g 20

13' Mean = 1.86
O L

E 0 N = 834.00

1.00 1.50 2.00 2.50 3.00 3.50 4.00
1.25 1.75 2.25 2.75 3.25 3.75

GFADCMP
Figure 4: Distribution of FAD General Functioning Scores

Parental acceptance-rejection was divided into two independent variables
including maternal and paternal scores from the same measure, the Parental Acceptance-
Rejection Questionnaire (PARQ) developed by Ronald Rohner. Seven items from the
neglect-indifference scale and the entire warmth-affection scale were reverse scored for
the purpose of making all responses consistent with the scoring (high indicative of
rejection; low indicative of acceptance). A total composite score was derived from the
data as well as composites from the four subscales--parental warmth, hostility, neglect,
and undifferentiated rejection for both mother and father.

The total raw composite score can range from 60 to 240 (midpoint = 150) using
60 items with Likert responses ranging from one to four. A score of 60 represents the
maximum perceived acceptance score and a score of 240 represents the maximum
perceived rejection score. This raw composite is useful for interpretation, however the

PARQ scoring manual recommends transfomiing the raw scores into Z-scores for more

74

accurate accounting of the data variation, considering the four sub-scales different sizes.
Z-scores were used to standardize the four subscales prior to summing them to form a
total composite Z-score.

With a range of possible raw scores ranging from 60-240 and 150 indicating the
recommended cut-off point, 95.6% of the sample scored 149 or below on the measure,
indicating a high rate of maternal acceptance. Nearly ﬁve percent (4.4%) of the sample
indicated maternal rejection with a raw composite score on the MPARQ of 150 or
greater. Of the total sample of 834, there was one missing MPARQ. This individual
completed the FPARQ and was therefore included in the data’s analysis. Six individuals
who did not complete either the FPARQ or the MPARQ were omitted from analysis.
These individuals overlapped with those who chose not to complete the EAT—26.

Table 6: Maternal Parental Acceptance-Rejection Data: Raw and Z-scores

 

 

 

 

 

Maternal Maternal Warmth- Aggression- Neglect- Undifferentiated
Acceptance- Acceptance- Affection Hostility Indifference Rejection
Rejection Rejection Sub-scale Sub-scale Sub-scale Sub-scale
(Raw Score) (Z-score) (Z-score) (Z-score) (Z-score) (Z-score)
Mean 89.906 -. 1664 -.0264 -.0543 -.0270 -.0570
Median 80.000 -1 .4217 -.4758 -.4466 -.2754 ~. 1870
Std. Dev. 26.4849 3.5594 .9770 1.0050 .9961 .9994
Minimum 60.00 -4.36 -l.2047 -1.1845 -l.1209 -1.4110
Maximum 225.00 17.66 5.3553 4.3501 4.7970 4.7089

 

 

 

 

 

 

Frequency distributions for both the raw score composite of the MPARQ and the
Z-score composite of the MPARQ to assess for their relative similarity. Although Z-
scores were calculated to standardize any variability among the four sub-scales, it was
assumed that the histograms for the raw and Z-score composites would be similar. Charts
3 and 4 show these distributions with normal curves included, evidencing the expected

similarity of the distributions.

75

300

200‘

Frequency

Frequency

 

 

Std. Dev = 26.48
Mean = 89.9
N = 833.00

 

 

/ /// / 9999
(90.09% 92200 (’0 Jo’od’oao’o ‘90 o;%0%0’o_0"oo~’oo

%0)0‘0 Wooooouo

MPARQCMP
Figure 5: Histogram for Raw Composite of MPARQ

 

200

 

 

 

Std.Dev=3.56
Mean=-.2
N=833.00
o a 7 6‘ (P /
Togo 0 0 0 0 '0 Qoeo'o‘od’o
ZMARQCMP

Figure 6: Histogram for Z—score Composite of MPARQ

76

Reliability measures to assess the internal consistency of the MPARQ and its four
scales yielded satisfactory results. The raw scale scores were used in calculated these
alphas. The 60-item MPARQ yielded an alpha of .9693. The four sub-scales internal
consistency scores are as follows: Warmth-Affection or = .9490 with 20 items;

Aggression-Hostility or = .9234 with 15 items; Neglect-Indifference or = .9065 with 15

items; Rejected-Undifferentiated or = .7861 with 10 items.

The paternal acceptance—rejection (FPARQ) scores showed greater dispersion
than the maternal acceptance-rejection (MPARQ) scores, however this variability was
accounted for by a small number of participants scoring higher on the FPARQ. Whereas
with the MPARQ 95.6% of the sample scored 149 or below, 90.4% of the sample scored
below 149 on the FPARQ. In looking closer at this data, two subjects scored above 220
on the MPARQ and three subjects scored above 220 on the FPARQ. Therefore the mean
of the data was affected by the highest score of 234 on the FPARQ. Nearly ten percent of
the sample indicated perceived paternal rejection, scoring 150 or greater on the FPARQ
as compared to the 4.4 percent of the sample indicating perceived maternal rejection. Of
the total sample of 834, there were nine missing cases for the FPARQ resulting in a
sample of 825 to be analyzed.

Table 7: Paternal Parental Acceptance-Rejection Data: Raw and Z—scores

 

 

 

 

 

Paternal Paternal Warmth- Aggression- Neglect- Undifferentiated
Acceptance- Acceptance- Affection Hostility Indifference Rejection
Rejection Rejection Sub-scale Suboscale Sub-scale Sub-scale
(Raw Score) (Zoscore) (Z-score) (Z-score) (Z-score) (Z-score)
Mean 96.3988 -.2384 -.0334 -.0920 -.0243 -.O887
Median 85.000 -1 .3645 -.4064 -.3997 -.3233 -.3423
Std. Dev. 33.3901 3.4376 1.0105 .9685 1.0090 .9923
Minimum 60.00 -4.13 -.90477 -.93577 -1. 1559 -l .4152
Maximum 234.00 14.27 3.3673 3.8885 3.9440 4.1282

 

 

 

 

 

 

77

As was evidenced with the MPARQ, the FPARQ’s two total composite scores,

both raw and Z—scores, are quite similar. For both the MPARQ and the FPARQ, Z-scores
will be used for subsequent analysis addressing the study’s research questions. Raw

scores are more useful however in describing the response cut-offs and interpreting their

meaning with the identiﬁed cut—off score of 150.

 

200

 

 

 

>5

8 Std. Dev = 33.39
0)

a. Mean = 96.4

[E N = 825.00

cane/09040/oa/ eee

FPARQCMP
Figure 7: Histogram for Raw Composite of FPARQ

78

 

200

 

 

 

>5

8 Std. Dev = 3.44
D

:0" Mean = -.2

1E N = 825.00

V‘J‘ ‘x0./.<?~??J.‘°f?d.’-?///”
0030.00 00 o o oo oo 000408020270

ZFARQCMP
Figure 8: Histogram for Z-score Composite of FPARQ

Reliability measures to assess the internal consistency of the FPARQ and its four
scales also yielded satisfactory results. The raw scale scores were used in calculated these
alphas. The 60-item FPARQ yielded an alpha of .9708. The four sub-scales internal
consistency scores are as follows: Warmth-Affection or = .9729 with 20 items;

Aggression-Hostility or = .9366 with 15 items; Neglect-Indifference 01 = .8831 with 15

items; Rejected-Undifferentiated or = .8163 with 10 items.
Hypotheses
Five primary hypotheses are addressed using the data in this study. Two
hypotheses aim to clarify the differences between the disordered and non-disordered
eating groups based on the independent variables of parental acceptance-rejection and
family functioning. Independent sample t-tests were used to address the two hypotheses
assessing for the level of difference between the disordered and non-disordered eating

groups based on the independent variables of parental rejection and family functioning.

79

Multiple regression analysis will be used to clarify the predictive relationship between the
independent variables and the dependent variable, eating attitudes and behaviors as
measured by the EAT-26.

Hymthesis l

The ﬁrst hypothesis states that there will be signiﬁcant differences between
disordered and non-disordered eating groups based on the variables of maternal and
paternal rejection. Both maternal and paternal acceptance-rejection variables were used
with the grouping variable of EATCUT, the dichotomous conversion of the continuous
dependent variable from the EAT-26. The EATCUT variable includes two groups with
1.00 code representing non-disordered scores measured as < 20 and 2.00 code
representing disordered eating scores on the EAT-26 of 20 or <. Both raw scores and Z-
scores are included in Table 8 for interpretation and comparison purposes. Statistics will
be reported for hypothesis 1 using Z-scores as recommended by the scoring manual for
the Parental Acceptance-Rejection Questionnaire.

The assumption of homogeneity of variance was not met in the analysis as
evidenced by p-values not exceeding the .05 level. One likely contributor to this unmet
assumption is the unequal sample sizes, n=659 and n=651 for group 1.00 (ZMARQCMP
and ZFARQCMP) and n=175 for both variable groups coded 2.00. Due to heterogeneity
being presence in the samples’ variances, t-test analysis reported in Table 9 represents
equal variances not assumed using SPSS.

The null hypothesis states that the disordered eating and non—disordered eating

groups will not be signiﬁcantly different based on the testing variable of maternal

80

acceptance-rejection (u1=u2). T-test results indicate that this null hypothesis be rejected
as a signiﬁcant difference exists between the groups’ means (t=—3.691, p=.001).

G*Power software program (Buchner, Erdfelder, & Paul, 1997) was used to
conduct post hoc power analyses in order to identify effect sizes for both t-tests and
regression analyses. For power analysis of the t-test based on the means of maternal
acceptance-rejection between the disordered eating (n=175, mean=96.63) and non-
disordered eating group (n=658, mean=88. 12), resulted in an effect size of d=0.33.
Therefore, although the groups were indicated as signiﬁcantly different using probability
values, the incorporation of effect size makes the analysis interpretation more tenuous. A
d-value of 0.33 is considered relatively small according tO Cohen ( 1977). Cohen (1977)
identiﬁes effect size conventions for t-tests to include small=0.20, medium=0.50, and
large=0.80.

Paternal acceptance-rejection was then inserted as the test variable. As was found
with the maternal variable, the null hypothesis is rejected (1.11.1412) with signiﬁcant
differences between the disordered eating and non-disordered eating means using
paternal acceptance-rejection (t=-3.415, p=.001). Power analysis for the paternal
acceptance-rejection t-test between disordered eating (n=175, mean=103.99) and non-
disordered eating group (n=650, mean: 94.35), resulted in an effect size of d=0.30.
Again, although the p—value indicates statistical signiﬁcance in the difference between the

two groups, the effect size is notably modest.

81

Table 8:
Group Statistics for Parental Acceptance-Rejection Variables included in T-tests

 

 

 

 

EAT-26 Cut-off N Mean Std. Deviation
Maternal
Acceptance- l.00 658 ' 88.1170 25.5708
Rejection 2.00 175 96.6343 28.7665
Raw Composite
Maternal
Acceptance- 1.00 658 -.4153 3.4241
Rejection 2.00 175 .7780 3.8960
Z-score Composite
Paternal
Acceptance- 1 .00 650 94.3554 32.0881
Rejection 2.00 175 103.9886 36.9563
Raw Composite
Paternal
Acceptance- 1 .00 650 -.4697 3.2777
Rejection 2.00 175 .6207 3.8662
Z-score Composite

 

 

 

 

Note. EAT-26 Cut-off: 1.00: non-disordered (score <20 on EAT-26); 2.00=disordered eating (score 20 or
< on EAT-26).

 

 

 

 

 

Table 9:
Homogeneity of Variance and T-test Analyses for Parental Acceptance-Rejection
Variables -
Levene’s Test for t-test for Equality of Means
Equality of Variances
F Sig. t df Sig.
(2-tai1ed)

Maternal Acceptance-
Rejection
Raw Composite 1 1.76 .001 -3.561 251.870 .000
Equal variances not
assumed
Maternal Acceptance-
Rejection
Z-score Composite 11.68 .001 -3.691 250.038 .000
Equal variances not
assumed
Paternal Acceptance-
Rejection
Raw Composite 11.38 .001 —3. 144 249.051 .002
Equal variances not
assumed
Paternal Acceptance-
Rejection
Z-score Composite 13.85 .000 -3.415 245 .391 .001
Equal variances not
assumed

 

 

 

 

 

82

Hypothesis 2

The second hypothesis states that there will be signiﬁcant differences between
disordered and non-disordered eating groups based on the variable of general family
functioning. The general functioning scale from the Family Assessment Device was used
as the grouping variable for the test variable, EATCUT. As stated above, EATCUT is a

categorical variable including two groups, non-disordered eating (1.00) and disordered

 

eating (2.00).
Table 10:
Group Statistics for General Family Functioning Variable included in T-test
l EATCUT l N I Mean 1 Std. Deviation
General Family 1.00 659 1.8180 .5589
Functioning 2.00 175 2.0157 .6300

Note. EAT-26 Cut-off: 1.00: non-disordered (score <20 on EAT-26); 2.00=disordered eating (score 20 or
< on EAT-26).

 

 

Table 11:
Homogeneity of Variance and T-test Analyses for General Family Functioning Variable
Levene’s Test for t-test for Equality of Means
Equality of Variances
F Sig. t df Sig.
(2-tailed)

General
Family
Functioning
Equal 3.842 0.51 -4.047 832 .000
variances
assumed

 

 

 

 

 

The null hypothesis for this research question states that the disordered eating and
non-disordered eating groups will not be signiﬁcantly different based on the testing
variable of General Family Functioning (111:112). The signiﬁcance level for both t-values
is .001 with t=-4.047 (equality assumption met) and t=-3.775 (equality assumption

unmet). Based on these statistics the null hypothesis is rejected. A signiﬁcant difference

83

exists between the two groups’ means based on general family functioning. Power
analysis for the t-test comparing the means Of the disordered eating (n=175, mean=2.016)
and non-disordered eating group (n=659, mean=1.818) based on the general family
functioning variable shows an effect size of d=0.358. Cohen (1977) identiﬁes effect size
conventions for t-tests to include small=0.20, medium=0.50, and large=0.80.

Hypotheses 3 and 4

Hypotheses three and four state the inﬂuences of the independent variables,
maternal acceptance-rejection, paternal acceptance-rejection, and unhealthy family
functioning on the outcome variable, eating attitudes and behaviors.

The ﬁrst step in this analysis process included checking the presence of
collinearity between these variables. The issue of multicollinearity is important to
recognize and address when conducting multiple regression analyses. Multicollinearity is
said to be present when two or more independent variables are strongly correlated with
one another (McClendon, 1994; Schroeder, Sjoquist, & Stephan, 1986). Multiple
regression methods assume the absence of perfect multicollinearity, whereby none of the
independent variables be perfectly correlated with another independent variable. It is
likely that some degree of multicollinearly is present in all regression analyses, since few
independent variables are totally uncorrelated (Schroeder, Sjoquist, & Stephan, 1986). No
single statistical test can determine whether or not the collinearity in the data is a problem
(Schroeder, Sjoquist, & Stephan, 1986). Schroeder, Sjoquist, and Stephan (1986) suggest
multicollinearity can be assessed by looking for high correlation coefﬁcients between the

variables to be included in the regression equation. Lewis-Beck (1980) recommends that

84

correlation coefﬁcients of .8 or larger are particularly concerning, and often useful in

diagnosing collinearity problems.

Table 12 shows the correlative relationships between maternal rejection

(ZMARQCMP), paternal rejection (ZFARQCMP), and general family functioning

(GFADCMP). As would be expected with social science instruments measuring various

aspects of family dynamics, all of the independent variables co-vary to some degree. The

strongest correlation is between the GFAD and the ZMARQCMP (1:.639). This

association is noted when interpreting the multiple regression model whereby the GFAD

coefﬁcient is dropped from the model using the step-wise method and ZMARQCMP is

retained. Therefore, although correlations exist between the independent variables, they

are considered modest enough (near r=.500) to accept in the current analysis.

Table 12: Correlation Data for the 3 Primary Independent Variables

 

 

 

 

 

 

Ge Maternal Paternal
neral
Family Acceptaonce- Acceptance-
Functioning Rejection Rejection
(Z-score) (Z-score)
General Family
Functioning
Pearson Correlation 1.000 .639** .562**
Sig. (2-tailed) .000 .000
N 834 833 825
Maternal Acceptance-
Rejection (Z-score)
Pearson Correlation .639** 1.000 .528**
Sig. (2-tailed) .000 .000
N 833 833 824
Paternal Acceptance-
Rejection (Z-score)
Pearson Correlation .562** .528** 1.000
Sig. (2—tailed) .000 .000
N 825 824 825

** = Correlation is signiﬁcant at the 0.01 level (2-tailed).

85

The three primary independent variables listed above are included in the small
regression model used in the analysis. An expanded model has also been incorporated to
reﬂect the four sub-scales present in the Parental Rejection-Acceptance Questionnaire as
well as the seven scales represented in the Family Assessment Device. The abbreviated
model includes three predictor variables and the expanded model employs ﬁfteen
predictor variables. Findings from these two models will be reported in an effort to gain
further clariﬁcation regarding variable relationships. Correlations between the ﬁfteen

independent variables used in the expanded regression model is provided below in Table

 

 

 

 

17.
Abbreviated Model Variables Expanded Model Variables
General Family Functioning Family Problem-Solving
Maternal Acceptance-Rejection Composite Family Communication
Paternal Acceptance-Rejection Composite Family Roles

 

Family Affective Responsiveness
Family Affective Involvement
Family Behavior Control
General Family Functioning
Maternal Warmth-Affection
Paternal Warmth-Affection
Maternal Aggression-Hostility
Paternal Aggression-Hostility
Maternal Neglect-Indifference
Paternal Neglect-Indifference
Maternal Undifferentiated Rejection
Paternal Undifferentiated Rejection

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Figure 9: Abbreviated and Expanded Regression Models
The normality Of residuals assumption was assessed. This normality assumption
was not met when using EATCOMP as the dependent variable. The scatterplot and
normal p-p plot evidence the non-normality of residuals in Figures 10 and 11. However,
when converting the outcome variable to a categorical variable (EATLOG), the

assumption was met as evidenced by Figures 12 and 13. In weighing the value in meeting

86

this assumption versus retaining a continuous variable as the outcome measure, the
project seems to gain more value more from linear regression than from logistic or binary

regression analyses. Therefore it is noted that non-normality of residuals exists within the

dataset.

 

200

 

 

>5

8 Std. Dev = .99
o

g Mean = 0.00
E N = 825.00

 

/€€JJY

Regression Standardized Residual

Figure 10: Histogram of Residuals using EATCOMP as Dependent Variable

87

 

 

 

 

'8

5:

E .25 ‘

L)

we

8

U

8

55 0.00 _ _ .

0.00 .25 .50 .75
Observed Cum Prob

Figure 11: Normal P-P Plot of Residuals using EATCOMP as Dependent Variable

Frequency

Figure 12

 

 

23‘4‘4‘-
erected

9"
4.
43
at.

4a
\1

Regression Standardized Residual

Std. Dev = .99
Mean = 0.00
N = 825.00

 

 

: Histogram of Residuals using EATLOG as Dependent Variable

88

 

 

 

Expected Cum Prob

 

 

0.00 .25 .50 .75 1.00

Observed Cum Prob
Figure 13: Normal P-P Plot Of Residuals using EATLOG as Dependent Variable
The enter method in multiple regression analysis was used to examine the

abbreviated data model with the three independent variables of maternal rejection,
paternal rejection, and general family functioning. This model accounts for a modest 3.9
percent of the total variance in the dependent variable of eating attitudes and behaviors
(R’=.039, Adj. R2=.036). Maternal acceptance-rejection is the only statistically signiﬁcant
predictor in the model. In comparing the effects of the standardized coefﬁcients in the
model, maternal acceptance-rejection has nearly twice the effect of paternal acceptance-
rejection (8:070) and four times the effect of family functioning (13:.034). For every
standard unit change in maternal acceptance-rejection, there will be a .126 increase in
eating attitudes and behaviors. An increase in the dependent variable as measured by the

EAT-26 signiﬁes greater disordered eating. See Table 13 below for regression coefﬁcient

data.

89

The null hypothesis is rejected based on a p-value of .001, indicating that the
model can be inferred to the population. The F~value is 11.203 for the model and is
statistically signiﬁcant. See Table 14 below for ANOVA statistics.

For power analyses of multiple regression, the effect size index P was used,
reﬂecting the proportion of variance accounted for by some source in the population
relative to the residual variance proportion (Buchner, Erdfelder, & Paul, 1997).
Regression analysis effect size conventions include small=0.02, medium=0. 15, and
large=0.35. Results from the regression power analysis for the abbreviated regression
model including three predictors (maternal acceptance-rejection, paternal acceptance-
rejection, and general family functioning) with a sample of 823 equaled 0.041. This is a
relatively small effect size according to Cohen (1977).

Table 13: Summary of Regression Analysis using the Enter Method for the Study’s
Abbreviated Model Predicting Eating Attitudes and Behaviors (N=823)

 

 

 

 

Variable Unstandardized Coefﬁcients Standardized Beta
B Std. Error

Gem?“ llamy .693 .986 .034
Functromg
Maternal
Acceptance- .425 . 157 . 126**
Rejection (Z-score)
Paternal
Acceptance- .240 . 148 .070
Rejection (Z-score)

 

 

 

Note. R’=.O39. Adj. R2=.036. **p<.01.

Table 14: Summary of ANOVA for Study’s Abbreviated Model using Enter Method

 

 

 

Sum of Squares df F
Regression 4579. 166 3 1 1.203* *
Residual 1 1 1720.0 820
Total 1 16299.2 823

 

 

 

Note. "p<.001

90

The step-wise method was also incorporated into the regression analysis of the
abbreviated model yielding results recommending the elimination of the GFAD variable
from the model. Two models were derived using the step-wise method. The ﬁrst model
entered only the maternal rejection variable, omitting both paternal rejection and family
functioning independent variables. In this case an R2 of .034 was achieved (Adj.=.033).
As compared to the enter method model used above, the standardized regression
coefﬁcient value for maternal rejection increased from 13:. 126 to 0:185. Additionally,
the F-value for this model is larger (29.114) than both the F—value for the enter method
abbreviated model (F=11.203) above and the second step—wise model (F=16.568) to be
discussed next.

The second model incorporated into the step—wise regression analysis included the
variables of maternal rejection and paternal rejection, again eliminating family
functioning as a valid predictor variable. The tolerance value, a collinearity statistic, may
provide some explanation why the GFAD predictor variable was eliminated from both
models. The tolerance value ranges from zero to one with lower values indicating high
multicollinearity. Tolerance is deﬁned as the proportion of the independent variables’
variance not shared with other independent variables. The tolerance value for GFAD in
model 1 of the step-wise analysis was .578 and .512 for model 2. Previous correlations
run among the independent variables indicated a moderate correlation between the
maternal rejection and family functioning variables (1:.639). It is likely that SPSS
dropped the family functioning variable from the model in part due to this collinearity

problem.

91

The R2 value for model 2 rose slightly from .034 in model 1 (ZMARQCMP only)
to .039 including both parental rejection variables. Interestingly when comparing the
second model in step-wise regression to the previous enter method, the same R2 values
are attained (.039), however the step-wise method achieves this level without the family
functioning variable present in the abbreviated enter model above. Therefore it may be
concluded that family functioning (GFAD) is not a variable worthy of remaining in the
model to predict disordered eating attitudes and behaviors.

Table 15: Summary of Regression Analysis using the Step-wise Method for the Study’s
Abbreviated Model Predicting Eating Attitudes and Behaviors (N=823)

 

 

 

Model Unstandardized Coefficients Standardized Beta
B Std. Error
1 Maternal
Acceptance- .622 .l 15 .185**
Rejection
2 Maternal
Acceptance- .480 .136 .l43**
Rejection
Paternal
Acceptance- .276 .139 080*
Rejection

 

 

 

Note. Model I (ZMARQCMP only). R2:.034. Adj. 122:.033. Model 2 (ZMARQCMP & ZFARQCMP).
R2=.039. Adj. 122:.036. *p<.05. **p<.001.

 

 

Table 16:
Summary of ANOVA for Study’s Abbreviated Model using Step-wise Method
Model Sum of Squares df F
1 Regression 3978.280 1 29.1 14**
Residual 1 12320.9 822
Total 1 16299.2 823
2 Regression 451 1.746 2 16.568**
Residual 1 1 1787.4 821
Total 1 16299.2 823

 

 

 

Note. Model 1 (ZMARQCMP only). Model 2 ZMARQCMP & ZFARQCMP). **p<.001

In addition to the abbreviated regression analyses used above, including the three

primary independent variables of maternal rejection, paternal rejection, and family

92

functioning, an expanded model was also used in order to clarify the inﬂuence of
particular predictor variables on the outcome variable. The expanded model removes the
PARQ total composite scores for both mother and father, replacing them with the four
separate sub-scale scores that are combined to create the total composite. The expanded
model also includes the additional six sub-scales from the Family Assessment Device,
beyond the general functioning score included in the abbreviated model. A total of
ﬁfteen predictor variables are therefore present in the expanded model as compared to the
three in the abbreviated model. It is noted that the r2 value will increase with the addition
of variables. The purpose of the expanded model was not to achieve a greater variance
accounted for, but rather to specify the most inﬂuential coefﬁcients on the outcome
variable.

Multiple regression analyses using the expanded model indicate speciﬁcations not
available through the abbreviated model because of the reliance on total composite
scores. The expanded model includes all seven of the Family Assessment Device sub-
scales as well as the four sub-scales within the Parental Acceptance-Rejection
Questionnaire for both mother and father, totaling ﬁfteen independent variables. The four
sub-scales for the Parental Acceptance-Rejection Questionnaire consist of
Warmth/Affection (WA), Aggression/Hostility (AH), Neglect/Indifference (N1), and
Undifferentiated Rejection (RU). The seven sub-scales for the Family Assessment Device
include general functioning, communication, roles, problem solving, affective
involvement, affectiveness responsiveness, and behavior control.

Table 17 outlines Pearson correlation coefﬁcients for the 15 independent variables

included in the model to check for collinearity. The majority of the correlations range

93

from .2 to .6. The highest correlations are shown between general family functioning and
problem solving (r=.777), affective responsiveness and general family functioning
(r=.789), maternal warmth and maternal neglect (r=.845), paternal warmth and paternal
neglect (r=.854) maternal aggression/hostility and maternal rejection (r=.859), and
paternal aggression/hostility and paternal rejection (r=.878). As was highlighted above,
the four sub-scales Of the Parental Acceptance-Rejection Questionnaire (warmth and
neglect; aggression and rejection) could likely be scaled down to two scales based on the

high correlations found in this study.

94

Table 17:
Correlation Malrlx for Study’s Independent Variables

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

.080. 0033 roam >33 >23. Worms mos. 2.032 mayo“ 2032 332 2.032 “~52
was: . mm . .302. 033. “:33. 532:. <<m::. >mm... >mm= Zomr z _.
a. .6883 .30 .mem .80 em» .86 .3 .90 .e»e .80 are. LS .eS
moi: 2 8e 8e was was was mos m8 m»m m8 m8 8» Eu
0633 68.8: a .48 .me m8 .ma .3» how .86 .84 new ads .emo awe
. 2 8a was 8a was 8e 8e 8m m8 m8 m»m m8 m»m
noose: . .mem mum m3 .9 » a: .3 u now .08 .85 .8» 62 .emn
moon _2 8e 8e ., 8e 8.» mos 8a 8a Sm mom a»... mum m»m
on. 68.83 .QO .68 was .96 Ha .mmm .98 4% .3... .e3
36 . 2 was as was . mos was was 8». m8 m8 m8 m8 m»m
on. house? e1.m» a; .2» .mmm .2: a1 and .»d hm .ﬂllemlel .98
.952 2 was 8e 8e mus 8e 8e 8m m»m m8 Sm mum m8
o as noose: .emm .8» a: s8 .3... at: 68 .»8 .8» .»»m and OS
0623. 2 8e 81s 8e 8.. 8... 8e m8 m8 m8 m8 80 m8
moo. 666.83 .3 mom so use 6.5 .9: .20 .8 68 .Sm .28. So
352. 2 was man 8a 8e was was was m8 m8 m»m m8 m»m
III] II
2.28. 68.83 as .eom emu .mmm saw .88 as an: .mmm .me .meu .e»m
263... z mMeI m8 m8 m8 m8 m8 m8 . m»s mom m»e m8 m»e
maﬁa: noose: . e»e .83 .08 .30 .»3 .»mo .80 an: .»8 as .8» .mme
5.63. .2 m8 m8 m»m m8 Em Em Em m»e m»e m»m m»a 8m
and nausea . use 88 .86 to to .8» .80 .mmm .»om .m: .3» ad
> a. 2 m8 m8 m8 m8 m8 m8 8m 8m we. 8.» m8 m»a
$56.. .0882: Sm .wme .mwm ﬁm .eom .mmm Sum .80 .90 .m: .to .mou
>98 2 gm 3m m»m m8 m8 Sm 8m Se Em Se 8e 3m
:2 2 noose: .elw...\l .80 do a: ems .3». .85 .mem .8» .3» .to at.
E8 .2 8w he m8 8». 8». mum m8 m8 3.. 8» m»... 8e
nose. nausea... .eBlA labs lam 1.31 teem .mem .801 N» was Mela: IE1 the“
26 _. 2 m8 m»m 8m m»m Em m»m mm. a»... 8m . m»... m»m Es .
:93: 68.83 .80 .88 .emo 8m 82 .w»m .moe . at .»S .86 .69 .8» a;
ma. 2 m8 m8 m8 wow 8m 86 m8 m8 m».. 86 m»e 8m Es
use noose? .»H . o a: .63 .8» . .i.».»u em». .8» .mom .emw Sm .eom 88
E. 2 8m Em 8m m8 Em m8 m8 8e Sm a? m»m m»e Sm
Zea”>:no:nE:5mEdm8:£8m=<am=508:m_muéwE60mb0f

 

 

 

 

 

 

2.032 amazon
36.. mo._.
.wmo .mum
mam mum
.wma .weo
mom mmw
.36 b:
mum mmm
.wwm .wd
mum mmm
.83 a
mum mum
.wmm .mmu
mum mmm
.mon .58
mmm mmm
.mﬁ .wvm
mum mme
.me .mom
mm.» mum
.mmw .emw
m8 mma
.mon .mum
mm.» mum

I
.3» .aom
mum mm.»
1.31 .I...8ol
mm.“ mmm
.mmm
. . mm;
.30
mme

 

95

The predictive ability of the model is enhanced somewhat with the incorporation
of the new variables as evidenced by an R2 of .089 (Adj. R2=.072). The F-value for the
model decreased from 11.203 to 5.268 (however both are signiﬁcant at p=.001). Perhaps
most useful with the expanded model is the enhanced ability to clarify what factors
within the maternal acceptance-rejection variable contribute most to the prediction of
disordered eating as well as investigating further impact by additional Family Assessment
Device variables incorporated.

The effect size calculated for the expanded model using 15 predictors including
sub-scale variables from the three primary variables for a sample of 823 reached 1"
=0.098. Regression analysis effect size conventions include small=0.02, medium=0.15,
and large=0.35. The f2 value for this model approaches a medium effect Size.

According to the regression coefﬁcients in Table 18, maternal warmth/affection

(B=-.213**), family problem-solving (B=-.161**), and family behavior control

(B=.100*), are signiﬁcant in the speciﬁed model in predicting disordered eating.

96

Table 18: Summary of Regression Analysis using the Enter Method for the Study’s
Expanded Model Predicting Eating Attitudes and Behaviors (N=823)

Unstandardized Coefficients

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Variable B Std. Error Standardized Beta
Fm?" Prob'em 4.155 1.433 -.161**
SolvrnL
Fam'" . . -.588 1.409 -.025
Communication
Family Roles -.871 1.584 -.031
Fam'" imam .234 1.097 .012
Responsrveness
Family Affective
Involvement 1 .627 1' 184 '065
Fa'm'y Behav'or 2.760 1.286 100*
Control
Gene'la' Tam” 3.391 1.832 .165
Functioning
Maternal Warmth- **
Affection .271 .093 -.213
Paternal Warmth-

Affection —.0352 .059 -.042
Maternal

Aggression- . 140 .108 .096
Hostility

Paternal

Aggression- -.0274 . 101 -.002
Hostility

Maternal Neglect-

Indifference .201 .121 .119
Paternal Neglect-

Indifference .0705 .091 .058
Maternal

Undifferentiated . 169 . 183 .069
Rejection

Paternal

Undifferentiated .0989 . 17 1 .046
Rejection

Note. 112-..089. Adj. R’=.072. *p<.05. **p<.01.

 

 

97

 

Table 19: Summary of ANOVA for Study’s Expanded Model using Enter Method

 

 

 

Sum of Squares df F
Regression 10359.829 15 5.268**
Residual 1059393 808
Total 1 16299.2 823

 

 

 

Note. **p<.001

As a follow-up to the expanded regression model, where 15 independent variables
were used as predictors, another regression model was run incorporating only those
variables showing signiﬁcance in the expanded model. These signiﬁcant variables
included family problem-solving, family behavior control, and maternal warmth-
affection. The purpose for running the follow-up regression analysis to the expanded
model was to explore the potential impact of collinearity on the variables.

Table 20: Summary of Regression Analysis for the Study’s Follow-up Model

 

 

 

 

Variable Unstandardized Coefficients Standardized Beta
B Std. Error

Fa'm'. y Pmblem -2.037 1.075 -.080
Solvrng
mm" Behav'" 4.682 1.073 4.363**
Control
Maternal Warmth-
Affection .0766 .05 1 .061

 

 

 

Note. 112:.030. Adj. 112:.027. **p<.001.

Table 21: Summary of ANOVA for Study’s Follow-up Model using Enter Method

 

 

 

Sum of Squares df F
Regpession 3513.118 3 8.573**
Residual 1 13241 .0 829
Total 1 16754.2 832

 

 

 

Note. **p<.001
The only variable to be statistically signiﬁcant in this follow-up model was family
behavior control (B=4.363**). Another notable change between the expanded model and

the follow-up model is the sign change evident in the maternal warmth-affection variable.

98

The expanded model showed the maternal warmth-affection variable to have a negative
correlation coefﬁcient (B=-.213), indicating an unsubstantiated relationship with
disordered eating according to academic literature, whereby as maternal warmth
increased, so to did disordered eating. The follow-up model indicates the maternal
warmth-affection variable to not only change signs from negative to positive, but also
shows it as non-signiﬁcant in the model.

The effect size calculated for the follow-up model using 3 predictors including
family problem solving, family behavior control, and maternal warmth-affection for a
sample of 832 reached P =0.031, a small effect size according to Cohen (1977).

Now that the multiple regression analyses have been reported, it is appropriate to
address the hypotheses stated earlier in the chapter. Hypothesis 3 states the presence of
parental rejection predicts disordered eating symptomatology. The relative impact of
parental rejection (both paternal and maternal) in predicting disordered eating is shown to
be small, with an r2 value of .039 (F=16.568**). The null hypothesis is rejected as
statistical signiﬁcance is reached to support the prediction of both paternal and maternal
rejection account for 3.9% of the variance in the outcome.

Hypothesis 4 discusses the relative impact of general family functioning on eating
attitudes and behaviors. In both the abbreviated and expanded regression models, general
family functioning was not shown statistically signiﬁcant. When utilizing the step-wise
regression method, general family functioning was eliminated from the model altogether.

Other sub-scales from the Family Assessment Device, measuring different aspects
of family functioning were found to be signiﬁcant. Most notably, family problem-

solving and family behavior control emerged in the expanded regression model as

99

signiﬁcant. Family behavior control was the only variable in the follow—up model to
surface as signiﬁcant.
Hypothesis 5

The ﬁnal hypothesis aims to deﬁne the relative impact of maternal and paternal
rejection on disordered eating symptomatology. Essentially this hypothesis focuses on
which of the parental variables is more effective at predicting the outcome variable. It
seems sufﬁcient evidence exists in the current data set to support drawing the conclusion
that maternal acceptance-rejection is more effective overall. The maternal acceptance-
rejection variable was statistically signiﬁcant and showed the highest standardized
regression coefﬁcients in both abbreviated regression models (13:. 126 enter method and
0:.185 step-wise method). The step-wise regression method, using the abbreviated
variable set, factored out maternal acceptance-rejection as the primary inﬂuential variable
in the model, with paternal acceptance-rejection contributing additional variance in the
second step.

The expanded model showed a lack of statistical signiﬁcance for any of the
paternal variables. Further, this expanded model indicated the maternal variable of
warmth/affection was most predictive of disordered eating. However, the follow-up
model indicated that the maternal warmth-affection variable was no longer signiﬁcant
when added to the model with family behavior control and family problem solving.

Therefore, based on three of the four regression models run in this study, the
maternal acceptance—rejection is more effective at predicting disordered eating than

paternal acceptance-rejection. None of the regression models run in the analyses found

100

paternal acceptance-rejection to be more signiﬁcant as a predictor compared to maternal

acceptance-rejection.

101

CHAPTER 5
Discussion
Introduction

Bruch (1974, p. 34) states that a “preoccupation with food may appear as helpless,
dependent clinging to parents, or as hostile rejection of them.” A wealth of studies have
examined this “dependent clinging” phenomenon as described by Minuchin and
associates (1978) as enmeshed or having boundary difﬁculties. This particular study
looked at the second contention in Bruch’s quotation, focusing on the predictive
relationship between the perception of parental acceptance-rejection and perception of
family functioning and the presence of disordered eating symptoms.

Young women at risk for developing an eating disorder are operationalized for the
purpose of this study as those scoring 20 or greater on the Eating Attitudes Test. This at-
risk population makes up the most commonly seen disordered eating group, consisting of
less severe problems centering around weight preoccupation, body dissatisfaction, eating
rules, and compensatory behaviors (Schwitzer, Bergholz, Dore, & Salimi, 1998). Recent
studies indicate 6% Of undergraduate women struggle with diagnosable eating disorders,
whereas a larger 25%—40% of this population face moderate problems, not reaching full
diagnosis (Koszewski, Newell, & Higgins, 1990; Schwitzer, et al., 1998).

In previous research, the 50th percentile of a female group of university students
received a score of 6 on the EAT-26 with the mean score of the subjects being 9.9
(Garner et al., 1982). The current study data indicates the median score to be 7.0 with the
mean score being 11.87 for the female sample surveyed. The current study found 21

percent of the sample qualifying as disordered eating according to the cut—off established

102

by Garner on the EAT—26. Walther (1994) also found a large percentage of subjects
having EAT-26 scores above the established cut-off in a study of 354 female university
students (23.4%).
Overview of Findings

In comparing the disordered eating and non-disordered eating groups in this
cross-sectional study as differentiated by the EAT-26, it was found that statistically
signiﬁcant differences exist between the two groups based on the variables of parental
acceptance-rejection and family functioning. The disordered eating group perceived their
family to be signiﬁcantly more unhealthy in their functioning according to the general
functioning scale from the Family Assessment Device. The disordered eating group also
perceived both their mothers and fathers to be signiﬁcantly more rejecting than the non-
disordered eating group based on the Parental Acceptance-Rejection Questionnaire.

In considering the entire sample Of over 800 female college students, the variables
with the greatest impact on the dependent variable of disordered eating included maternal
acceptance-rejection, speciﬁcally maternal warmth/affection, family problem solving,

and family behavior control.

103

Conceptual Models

Original Conceptual Model

   

  
   
   
   
 
       
   
  
   
     
    

PERCEPTIO
OF PATERNAL
ACCEPTANCE-
BEJECTION

 

 

PERCEIVED
UNHEALTHY DISORDERED
FAMILY EATING

FUNCTIONING

 

 

 

 

PERCEPTION OF
MATERNAL
ACCEPTANCE-
REJECTION

Figure 2: Conceptual Model

The original conceptual model incorporated the three primary independent
variables, the dependent variable, and the relationships between all variables in the model
using directional and shaded arrows. The darker the arrow, the stronger the relationship
between the variables in the model. Therefore, according to the model presented above,
there was expected to be some collinearity between the independent variables, with the
correlation between maternal and paternal acceptance-rejection being stronger than
between family functioning and either of the parent acceptance-rejection variables.
Existing academic literature supported the presence of darker arrows in the model

directed from each of the independent variables to the dependent variable, disordered

104

eating. Using recent literature differentiating the effects of mother and father variables on
disordered eating, it was hypothesized that paternal acceptance-rejection would show a
stronger relationship with disordered eating than maternal acceptance—rejection.

Revised Conceptual Model

PERCEPTION
OF PATERNAL

 

 

 
     
 

 

 

 

ACCEPTANCE-
3...} REJECTION , _
5’7 A \
PERCEIVED
UNHEALTHY _ ..._ DISORDERED
FAMILY ‘ EATING
FUNCTIONING

 

 

 

 

 

v a;

PERCEPTION OF
MATERNAL
ACCEPTANCE-
REJECTION

 

Figure 14: Revised Conceptual Model
The revised conceptual model reﬂects ﬁndings presented in chapter 4. There were
correlations between all of the independent variables; however, in contrast to the original
conceptual model, the strongest correlation was between the family functioning and
maternal acceptance-rejection variables. Multiple regression analyses using the three
primary independent variables, general family functioning, maternal acceptance-
rejection, and paternal acceptance-rejection, led to the ﬁnding that the strongest

relationship exists between the maternal acceptance-rejection variable and disordered

105

eating. The expanded and follow-up regression models support the inclusion of family
functioning as an inﬂuential predictor when operationalizing it through problem-solving
and behavior control variables. Therefore, the independent variables in the conceptual
model above which serve as most inﬂuential include maternal acceptance and family
functioning. Paternal acceptance-rejection did not show signiﬁcant in any regression
model run, therefore its minimal relationship to the dependent variable is indicated with a
light gray arrow.
Evaluation of Research Questions

Five research questions were outlined for the current study based on relationships
between the independent variables of maternal acceptance-rejection, paternal acceptance-
rejection, and family functioning with the dependent variable, eating attitudes and
behaviors. These research questions will each be addressed with supporting data and
literature below.
Research question 1

Research question 1 asks, “Does parental acceptance-rejection differ between
individuals with disordered eating and those without?” The means for maternal
acceptance-rejection included 88.12 (Z-score=-.4153) for the non-disordered eating group
and 96.63 (Z—score=.7780) for the disordered eating group. The two groups were found to
be signiﬁcantly different, with the disordered-eating group showing higher rates of
maternal rejection. The means for paternal acceptance-rejection were 94.35 (Z-score=-
.4697) for the non-disordered eating group and 103.99 (Z-score=.6207) for the disordered
eating group. The two groups were also found to be signiﬁcantly different with the

disordered eating group indicating higher rates Of paternal rejection. Therefore, according

106

to the data in this study, disordered eating individuals show statistically signiﬁcant higher
levels of both maternal and paternal rejection in comparison to their non-disordered
eating counterparts. Hypothesis one is thereby supported, with both null hypotheses being
rejected with statically signiﬁcant differences existing between the two groups. However,
it is important to note that the data indicated a small effect size for both the maternal and
paternal t-tests.

Research question 2

Research question 2 asks, “Does family functioning differ between individuals
with disordered eating and those without?” The means for the general family functioning
scale derived from the Family Assessment Device were 1.82 (SD=.56) for the non-
disordered eating group and 2.02 (SD=.63) for the disordered eating group. The null
hypothesis is rejected at a probability level of .001, indicating these two groups are
signiﬁcantly different. Therefore hypothesis two is supported, which stated that the
disordered eating group would show signiﬁcantly higher levels of unhealthy general
family functioning than the non-disordered eating control group. Again, although
statistical signiﬁcance was found using a t—test analysis, the effect size was relatively
small at d=0.358, making the results more tenuous to interpret.

In a study looking at the relationship between Eating Disorder Inventory-2 scores
with Family Assessment Device scores, McGrane and Carr (2002) used t-tests to analyze
the data as was conducted in the current study. The research team found signiﬁcant
differences on the dimensions of general functioning, problem-solving, roles, and
affectiveness responsiveness between women in the eating disorder group versus the

normal controls. In contrast, Kent and Clopton (1988) examined a non-clinical college

107

sample and found no signiﬁcant differences reported between family relationship patterns
of bulimics and normals.
Reseaach question 3

Research question 3 asks, “Does the presence of parental rejection predict
disordered eating symptomatology?” Abbreviated regression models were used to
address this research question, including only the three primary variables (maternal
acceptance-rejection, paternal acceptance-rejection, and general family functioning).
Only the maternal acceptance-rejection composite variable was signiﬁcant in predicting
disordered eating attitudes and behaviors using the ﬁrst model. However, in step—wise
regression analysis, two models were pulled out as most predictive using the three
primary variables. The ﬁrst included only maternal rejection-acceptance, conﬁrming this
variable as most inﬂuential in the regression model. The second included both maternal
and paternal rejection-acceptance variables as signiﬁcant. A further breakdown Of the
analysis of maternal and paternal variables is provided in research question ﬁve below.

The relative impact of parental rejection (both paternal and maternal) in
predicting disordered eating is shown to be signiﬁcant but small, with an r2 value of .039
(F=16.568**). Overall, the data suggest that maternal acceptance-rejection is the most
inﬂuential variable in the model when considering the standardized regression
coefﬁcient. In comparing the effects of the standardized coefﬁcients in the abbreviated
model, maternal rejection has nearly twice the effect of paternal rejection and four times
the effect of general family functioning. Therefore, generally speaking, perceptions of

parental acceptance-rejection are related to the outcome of disordered eating. This

108

relationship however is explored further through research question ﬁve which speciﬁes
the inﬂuence of maternal versus paternal variables.

It was hypothesized that parental acceptance-rejection variables would contribute
signiﬁcantly to predicting disordered eating based on a number of studies. This
hypothesis was based, in part to a study recently published by Dominy, Johnson, and
Koch (2000) which found that women with binge eating disorder reported their father to
be more rejecting and less warm in comparison to mothers. Other academic literature
supporting this hypothesis included numerous studies using Rohner’s rejection construct
that found it to be signiﬁcantly correlated with other psychosocial problems such as drug
and alcohol abuse, schizophrenia, depression, conduct disorder, externalizing behaviors,
and delinquency (Khaleque & Rohner, 2002; Rohner & Nielsen, 1978). The majority of
these studies used t-tests to analyze if differences exist between treatment (unhealthy) and
control (healthy) groups, but did not run more complex analyses such as multiple
regression analyses or structural equation modeling. As evidenced through research
question one above, results of the current study could be interpreted as contributing to the
literature supporting the connection between parental rejection and psychosocial
problems. However, when looking at this relationship through regression analysis, a
correlational statistical test, it becomes more difﬁcult to interpret with conﬁdence. The
abbreviated models used in this study show parental acceptance-rejection (particularly
maternal acceptance-rejection) to be signiﬁcant in relationship to disordered eating,
however the variance accounted for in the outcome variable is relatively small.

The complexity of the outcome variable is critical to consider when thinking

about what sizes of r2 values would be expected. Disordered eating has been described as

109

a multi-factorial and multi-dimensional construct. The complexity of this variable is not
likely to be explained by any one primary predictor variable such as parental rejection.
Through a human ecological framework, Bronfenbrenner (1979) theorizes that
individuals live within multiple systems, interacting with millions of people, symbols,
and objects throughout their lives. In considering this framework, it is useful to ﬁnd that
nearly 4% of the disordered eating variance is accounted by parental rejection. However,
it is just as useful to recognize that many other variables, both within and outside of the
family, contribute to the onset and maintenance of the illness as well, such as peers, role
of media, genetics, life cycle stage, gender, and incidences of traumatic events to name a
handful in the research literature.
Research question 4

Research question 4 asks, “Does the perception of unhealthy family functioning
predict disordered eating symptomatology?” Of the three primary independent variables,
general family functioning showed the least impact relative to the outcome variable and
was dropped from the model when using step-wise regression techniques. The general
family functioning sub-scale was chosen as one of the primary independent variables for
the abbreviated model analysis because it includes various aspects of family functioning.
Therefore, during the ﬁrst stage of analysis, the investigator would have responded to this
research question by stating that unhealthy family functioning does not serve as an
inﬂuential variable in predicting disordered eating. However, in the second round of
analysis when an expanded regression model and follow-up model were used, different

results emerged when considering the variable of family functioning.

110

As mentioned earlier, the Family Assessment Device contains seven sub-scales.
All of these sub—scales were included in the expanded regression model to ﬁnd if other
sub-scales, beyond general family functioning were useful in predicting disordered
eating. Results from the expanded regression model indicated that two sub-scales from
the Family Assessment Device were signiﬁcant in predicting the outcome variable. These
included family problem-solving and family behavior control.

Family problem solving evidenced a negative relationship with disordered eating.
In other words, as family problem solving decreased, indicating greater ease in
developing solutions in the family, the likelihood of disordered eating increased. The
problem solving scale refers to the family’s abilities to resolve problems within and
outside the family at a level which maintains effective family functioning (Miller,
Epstein, Bishop, & Keitner, 1985). This ﬁnding is inconsistent with existing literature,
which has most often tended to highlight a positive relationship between the presence of
eating disorders and poor problem-solving, particularly in the cases of bulimia. Waller,
Calam, and Slade (1989) found women with bulimia were shown to rate their family as
having problem solving difﬁculties according to the FAD. Additionally, Waller (1994)
found the disordered eating symptom of bingeing to be positively associated with
perceptions of poor problem—solving. Two differences between these studies and the
present study include type of sample (clinical versus non-clinical) and sample sizes. First,
both studies cited here that found signiﬁcant positive relationships between poor problem
solving and the presence of eating disorders used clinical samples whereby participants
were recruited from eating disorder clinics. In contrast, the sample used in the present

study was a non-clinical group of college—aged women with no conﬁrmed eating disorder

11]

diagnoses. Second, the sample size of the current study included more than 800
individuals, whereby the other two studies cited here included an n=81 (Waller, 1994)
and n=78 (Waller, Calam, & Slade, 1989).

Another explanation for the interesting negative relationship between the
perception of healthy family problem solving and the presence of disordered eating may
be the likelihood for respondents with disordered eating to overstate the positive in their
family. The academic literature evidences differences between bulimic and anorexic
study participants in regards to how they report perceiving family interactions.
Speciﬁcally, women with bulimic symptoms tend to report less healthy interactions in
their families in comparison to women with anorexic (restricting) symptoms who tend to
present their family environment as more healthy (Scalf-Mclver & Thompson, 1989;
Waller, Calam, & Slade, 1989).

Family behavior control was the only signiﬁcant variable to emerge through the
follow-up regression model including the additional variables of problem-solving and
maternal warmth-affection. Family behavior control showed a positive relationship with
disordered eating. In other words, as family behavior control increased, indicating greater
control of family members behaviors, the likelihood of disordered eating also increased.
The behavior control scale refers to the ways in which a family expresses and maintains
standards of behavior for family members. This dimension of family functioning outlines
how a family handles behavior in three areas including physically dangerous situations,
situations that involve the meeting and expressing of psychobiological needs and drives,
and situations involving socializing behavior both inside and outside of the family

boundary (Epstein, Bishop, & Levine, 1978; Epstein, Bishop, Ryan, Miller, & Keitner,

112

1993). This ﬁnding is consistent with the literature as Hodges, Cochrange, and Brewerton
(1998) found individuals exhibiting binge-eating behaviors tended to rate their families as
having increased control than a group of controls. Waller, Calam, and Slade (1989) used
the Family Assessment Device to assess if women having both anorexia and bulimia
rated their families as more unhealthy. The research team found that women with
anorexia and bulimia rated their families as having more difﬁculty on the affective
involvement and behavior control scales of the Family Assessment Device (Waller,
Calam, & Slade, 1989). Also, a signiﬁcant positive relationship was found between the
variable of parental control and the presence of anorexia and bulimia in an adolescent
population according to Felker and Stivers (1994). In this study it was discovered that as
parental control increased, so too did eating disorder symptoms, mirroring the ﬁndings of
the current study.
Research question 5

Research question 5 asks, “Does maternal or paternal rejection serve as a better
predictor of disordered eating symptomatology?” The response to this research question
is partially embedded in question three above, where it was eluded that maternal
acceptance-rejection served as a more inﬂuential predictor in the regression model when
compared to paternal rejection. Maternal acceptance-rejection was twice as effective at
predicting the outcome variable as paternal acceptance-rejection according to the
regression coefﬁcients reported in the abbreviated regression model.

As was conducted with the family functioning variable in the analysis, the
parental acceptance-rejection composites were broken down into four sub-scale variables

(warmth-affection, aggression-hostility, neglect-indifference, and undifferentiated

113

rejection) for further exploration. These eight additional variables (four for maternal and
four for paternal) were included in the expanded regression model for the purpose of
deriving clariﬁcation on what sub-scales from the maternal variable were most
inﬂuential. The largest maternal regression coefﬁcient in the model was
warmth/affection. The ﬁnding of maternal warmth/affection being the most inﬂuential
variable is interesting primarily because of the presence of a negative correlation with the
outcome variable. According to the data, as maternal warmth/affection decreases,
indicating more overall warmth/affection because the scale is reverse coded for
consistency with the other PARQ sub-scales, disordered eating actually increases.
Therefore, the more warmth an individual perceives from her mother, the more likely she
is to exhibit disordered eating symptomatology. The relationship seems counterintuitive
and is perhaps the most surprising result from the study.

Maternal warmth/affection, as deﬁned by Rohner (1991), refers to mother-child
relationships where mothers are perceived to give love or affection, exhibit an interest in
her daughter’s activities and well-being, and Show approval of the child. One possible
way of explaining the relationship between high maternal warmth and increased
disordered eating is to think about the functionality of the symptoms. Numerous
clinicians, theorists, and researchers have postulated rationales for the presence of eating
disordered symptoms over the years. For instance, Walther (1994) states that eating
disorder symptoms might represent expressions of anger, ways of retaliating, or strategies
for competing in families with an overemphasis on achievement and thinness.
Williamson (1990) interpreted these symptoms as indirect forms of rebellion or “passive-

aggressive” behaviors. Still others postulate symptoms may serve the function of coping

114

with depression or anxiety, elicit attention and nurturance, or work to keep the family
(particularly the parents) together (Minuchin, Rosman, & Baker, 1978). One hypothesis
generated for this study is that eating disorder symptoms may be a way of seeking
acceptance and approval from parents, particularly in those families that value thinness,
appearance, and achievement. The results achieved may indicate that individuals perceive
greater maternal acceptance with the presence of disordered eating symptoms. In other
words, if a daughter believes that she would gain more acceptance from her mother if she
emulated more of what Western society deems beautiful for a female, achieving thinness
as one of these standards of beauty may act as a vehicle to soliciting increased maternal
warmth and affection. Although an interesting idea, much more research must be
conducted before this idea can be stated with any degree of certainty. Qualitative
interviews may be an ideal way to gather data from participants to either support or refute
this contention.

In further exploration of the relationship between maternal warmth and disordered
eating, a follow-up regression model was run with the three variables that emerged as
statistically signiﬁcant in the expanded model, maternal warmth-affection, family
problem-solving, and family behavior control. The follow-up regression model provided
insight into the unexpected relationship between maternal warmth and disordered eating
as the negative sign changed to a positive sign for the maternal warmth variable. This
sign change indicated that a different relationship was evidenced by the model, whereby
as maternal warmth decreased, disordered eating increased. This new relatiOnship is
likely more accurate as the follow—up model incorporated less variables and factored out

some of the collinearity problem amongst the independent variables. However, although

115

the sign changed, making the relationship between maternal warmth and disordered
eating consistent with existing literature, the relationship was non-signiﬁcant.

Due to the inconsistency in results between the various regression models run
within the current study, particularly with regards to the maternal acceptance-rejection
variable of maternal warmth-affection, caution is taken in interpreting, with any degree of
conﬁdence, what the data indicates. More research is needed to clarify the nature of the
relationship between maternal acceptance-rejection and disordered eating as
inconsistency was found in the results of this study’s analysis.

Limitations

As with any research study, limitations exist, and should ideally be identiﬁed
from the conception of a project as to make decisions about their impact on affecting the
results. A number of limitations within the current study’s design will be discussed
below. These limitations include sample, generalizability, use of self-report measures,
and undifferentiated eating disorder symptomatology.

.Sa1_np_le

In an effort to target a non-clinical sample fertile with disordered eating
symptomatology, a university setting was chosen. The inclusion of this speciﬁc female
population, as well as the utilization of a range of disordered eating symptoms versus
speciﬁc clinical diagnoses provides a wealth of data, but limits generalizability. The
ﬁndings in the study must be interpreted with caution as it is unsure if they are indicative
of sub-clinical levels of disordered eating, more common to extreme dieters rather than
speaking to the experiences and perspectives of a clinical sample with conﬁrmed eating

disorder diagnoses. Follow-up interviews to verify whether the hi gh-scoring college

116

students had a clinical eating disorder were not conducted. Further, because the high-
scoring groups contained some individuals who had relatively minor degrees of eating
disturbance, reported differences between groups may be conservative.

Beyond allowing for a range of eating disordered symptomatology and utilizing a
college sample, the sampling procedure utilized threatens generalizability. All of the
participants involved in the project were self-selected rather than randomly drawn from
the population. This may limit the results’ generalizability due to characteristics
potentially unique to individuals who self-select versus those who do not.

Self-report measures

Utilization Of self-report measures involving recall responses is a limitation to this
study as the Parental Acceptance-Rejection Questionnaire asked participants to think
back to when they were in late childhood to rate their responses. To off—set the limitation
inherent in using self-report measures, the investigator strived to incorporate measures
with particularly high validity and reliability as evidenced through standardization studies
of the instruments and other researchers using the measures over the past decade (Garner,
Olmsted, Bohr, & Garﬁnkel, 1982; Netemeyer & Williamson, 2001; Rohner, 1986, 1991;
Khaleque & Rohner, 2002; Epstein, Bishop, Ryan, Miller, and Keitner, 1993). Epstein,
Bishop, & Levin, 1978; Epstein, Baldwin, & Bishop, 1983; Kabacoff, Miller, Bishop,
Epstein & Keitner, 1990; Waller, 1994). However, it is possible that weaknesses in the
self-report instruments accounted for the small amount of variance explained in the
outcome in both the abbreviated and expanded regression models.

Useful information was solicited by examining participants’ perceptions of their

relationships with their parents and families. However, conclusions are limited to

117

participants’ subjective views, which may deviate in certain aspects from more objective
family interactional patterns obtained through Observation methods. One speciﬁc
potential problem with the use of self-report eating attitudes and behaviors rather than
conducting intensive clinical interviews to achieve the presence of eating disorders is the
possibility of denial preventing accurate recognition or reporting of disordered eating
symptoms. Due to the denial often characteristic of eating disorders, particularly anorexia
nervosa, participants may have been less aware of, or less willing to report unhealthy
aspects of family functioning or negative relationships with their parents. Vandereycken
and Vanderlinden (1983) state that the tendency to deny problems is not uncommon
among individuals with anorexia nervosa, and with increasing severity of the disorder, so
may the distortion of response pattern. This may have contributed to the small amount of
variance accounted for in the regression models presented.

Embedded within self—report measures is the aspect of social desirability. The
instruments used within this study target personal aspects of an individuals’ life including
their eating behaviors and attitudes, their perspective on their mother’s and father’s liking
of them, as well as their viewpoint of how their family functions. Even with assurance of
anonymity, participants may have answered the questionnaires more based on appearing
“healthy” than on providing true and accurate information.

Sequencing of the questionnaires

The data collection technique of having participants complete both mother and
father questionnaires within the same time period and sequentially may raise a question
of validity. Might the scale scores have changed if completed on separate occasions

without the tendency to compare immediately the perception of mother and father? It

118

may have been likely that more perceived maternal warmth magniﬁed the perception of
paternal neglect for participants. For instance, if an individual completed the maternal
warmth scale in a positive direction (as evidenced by the results), this comparison could
be made when completing the questionnaire regarding her relationship with her father. In
essence, the perception of these relationships is in fact, relative. Perhaps the more warmth
perceived from mother, the more it highlighted the neglect felt from the relationship with
father. Rather than seeing parents as a unit and responding with similar scores for each,
the respondents seemed to very much view their parents as independent entities in their
lives. The maternal warmth scores may have been inﬂated to account for the child coping
with rejection or neglect from her father. Or the paternal neglect scores may have been
inﬂated as the respondent compared her father’s engagement in her life with her
mother’s, which was already scored as high. In this case, there is no evidence of
regressing toward the mean for these two variables, but rather they seem to increase in

opposite directions as the level of disordered eating increases.

Undifferentiated eating disorder symptomatology

 

Interpretation of results and comparisons to previous research are also limited by
the inability to distinguish between subjects with speciﬁc eating disorder symptoms and
diagnoses, particularly anorexia nervosa, bulimia nervosa, and binge-eating disorder.
Certain expected relationships between disordered eating symptomatology and family
variables might have been insigniﬁcant or inconsistent due to the mixture of
indistinguishable symptoms with the EAT-26. The Eating Attitudes Test speciﬁes its
measure of anorexia nervosa symptomatology, however there are a number of questions

that seem to assess for bingeing and purging behaviors prevalent in bulimia nervosa.

119

Therefore, results from the present study may be limited in comparing its ﬁndings to
research whereby participants were separated based on symptoms, such as bingeing and
purging versus restricting.
Clinical implications

The sample utilized in the current study as well as the research aims call for two
different treatment modalities to be discussed here. First, because of the college-aged
sample, it is critical to consider their current environment, the university setting, when
incorporating clinical implications. Additionally however it is necessary to discuss
clinical implications regarding the content of the study which included perceptions of
family. Therefore, family therapy implications will be discussed in the context of this
section as well.
Treatment on the college campus

As many as 61% of college students have been found to display eating disordered
behaviors while not meeting the criteria for an eating disorder (Mintz & Betz, 1988).
Koszewski, Newell, and Higgins (1990) found that approximately one-quarter Of the
female undergraduates they surveyed felt that their eating was out of control. In the
current study, 21% of college females taking the EAT-26 scored above the clinical cut-
off, signifying problematic symptoms. As was found in a study conducted by Schwitzer,
Bergholz, Dore, and Salimi (1998), the items most highly endorsed on the Eating
Attitudes Test in the current study involved obsessive thoughts about food, fat on the
body, caloric intent, and exercise. The research team headed by Schwitzer noted that
many of the females surveyed did not meet full diagnostic criteria for an eating disorder,

but did however display serious symptoms highly correlated with the disorders.

120

Speciﬁc questions within the EAT-26 were analyzed to get a better sense of the
eating attitudes and behaviors of the target sample, which includes undergraduate female
college students. The most highly endorsed items focus on obsessions and include
preoccupation with food (30.9%), the calorie content in food (37.2%), the desire to be
thinner (43.3%), the thought of having fat on her body (36.2%), and burning up calories
during exercise (59.0%). Over half of the sample (53.7%), identify themselves as being
terriﬁed about being overweight. One hundred and six females ( 12.7%) reported
experiencing uncontrollable bingeing often, usually, or always. In contrast, Hesse-Biber,
Marino, and Watts-Roy (1999) report that between 60 and 80 percent of college women
engage in regular binge eating and other abnormal behaviors that fall short of diagnostic
criteria.

The EAT-26 assesses for vomiting in two questions, with the ﬁrst asking if the
behavior actually occurs and the other assessing the impulse, not necessarily the
behavior. Twenty-eight females (3.4%) reported vomiting after eating, whereas sixty-nine
(8.3%) reported having the impulse to vomit after eating. Koszewski, Newell, and
Higgins (1990) found 6% of a female undergraduate population surveyed reported

forcing themselves to vomit after meals or using laxatives as weight control measures.

121

Table 22: Frequencies for Speciﬁc EAT-26 Questions

 

 

 

 

 

 

 

 

 

 

EAT-26 Question Code" Frequency Percent
. . . 1.00 386 46.3
Am temﬁed about being overweight 2.00 448 53.7
. . . 1.00 576 69.1
Find myself preoccupied With food 200 258 309
Have gone on eating binges where I feel that I may 1.00 728 37-3
not be able to stop 2.00 106 12.7
. 1.00 524 62.8
Aware Of the calorie content of foods that I eat 2.00 310 37.2
. 1.00 806 96.6
Vomit after I have eaten 2.00 28 34
Am preoccupied with a desire to be thinner :% :2? 2g;
. . . . 1.00 342 41.0
Think about burning up calories when I exercrse 2.00 492 590
Am preoccupied with the thought of having fat on 1.00 532 63.8
my body 2.00 302 36.2
. 1.00 694 83.2
Feel that food controls my life 200 140 16.8
Have the impulse to vomit after meals :33 76695 98137

 

 

 

*Code: 1.00 = sometimes, rarely, or never ( non-disordered); 2.00 = often, usually, or
always (high risk for disordered eating)

It is recommended that treatment on a college campus consist of three
components as originally suggested by Schwitzer and colleagues (1998). The evidence
within the current study indicate that a signiﬁcant portion of young females are struggling
with disordered eating to a degree that likely interferes with interpersonal relationships,
physical, nutritional, and developmental growth, and cognitive abilities. The ﬁrst tier of
treatment includes prevention, followed by intermediate services and remedial treatment
(Schwitzer et al., 1998). The prevalence and range of symptoms found in the current
study suggest a thorough intervention framework is needed on college campuses today.
Prevention can be incorporated into introductory courses as well as both the dormitory

and cafeteria environments through nutritional and psychoeducation regarding the

122

beneﬁts of healthy eating and body image. Intermediate services, the second tier of the
intervention framework, targets sub-clinical disorders. These services might likely be
delivered through the campus health center, the athletic department, the Greek system,
and any campus mental health clinic through conducting psychoeducation groups aimed
at developing skills, shifting faulty cognitions, and incorporating new behaviors. The
third and ﬁnal tier of the framework involves more intensive individual, group, and
family therapeutic efforts. The campus mental health clinic may be an efﬁcient avenue
for the execution of such efforts. However, it should be noted that the inclusion of multi-
disciplined professionals at this level is ideal for comprehensive treatment. A common
multi-disciplinary team includes a therapist, nutritionist, physician, and at times a
psychiatrist.

Treatment within the family

The following section focuses on the current study’s ﬁnding that family
behavioral control was the most influential predictor of disordered eating. The variables
including age of the symptomatic individual, format of treatment, and goals of treatment
will be discussed as pertinent to the facilitation of family therapy. Each of these factors
will be explored through this lens of behavioral control and recommendations will be
made based on existing academic and clinical literature.

The ﬁrst therapeutic factor needing to be addressed by the family therapist is in
regards to the symptomatic individual’s age, which will help determine the appropriate
format of treatment. Recent research suggests that younger, adolescent individuals
displaying eating disorders beneﬁt greatly from conjoint family therapy (Minuchin,

Rosman, & Baker, 1978; Russell, Szmukler, Dare, & Eisler, 1987; Dare, Eisler, Russell,

123

& Szmukler, 1990; Eisler, Dare, Russell, Szmukler, Le Grange, & Dodge, 1997). Russell,
Szmukler, Dare, and Eisler (1987) conducted one of the few controlled trials comparing
family therapy with individual therapy. Their ﬁndings support that family therapy is more
effective than individual therapy in patients whose illness was not chronic and had begun
before the age of 19 years old. Young adults, such as those included in the current study’s
sample (ages 18-25), have been shown to beneﬁt from separated family therapy whereby
parents and the individual engage in separate counseling simultaneously. Some
researchers have provided evidence that a different form of family therapy (separated
rather than conjoint) may be better suited to their life cycle stage of launching from the
family system (LeGrange, Eisler, Dare, and Russell, 1992.

The eating disordered individual’s life cycle stage plays a key role in addressing
treatment. Considering the current study’s sample, the launching or individuation-
separation process is underway for the participants. Vanderlinden and Vandereycken
(1991), a team that has spent the vast majority of their research time studying eating
disorders in the context of family published a set of guidelines for a family therapeutic
approach to treating eating disorders. Within this approach it is recommended that
clinicians dedicate time and attention to facilitating the separation-individual process of
the symptomatic individual. Vanderlinden and Vandereycken (1991) found that patients
who actually leave the family’s home and start living independently function better than
patients who stay in their parent’s home.

Considering the ﬁnding that individuals with higher levels of disordered eating
perceive greater levels of behavior control, the launching process may be equally exciting

and terrifying. It is likely that their parents have managed many aspects of their lives up

124

through adolescence including ﬁnances, social schedules, and many major decisions such
as post high-school plans. The launching process necessarily decreases behavioral
control, shifting decision-making responsibility from the parents in the system to the
individual. In the case of an eating disorder, the individual is ﬁrst placed in charge of her
symptoms and then is encouraged to manage other areas of her life. Not only is this a
great shift for the young adult, but equally for the parents who may become extremely
anxious with their new experience of feeling out of control. At this stage in treatment,
Vanderlinden and Vandereycken (1991) advise the therapist to predict relapse, as the
family may take the cue of returning symptoms to mean they are once again needed to
control various aspects of their daughter’s life. ’

In combining the signiﬁcant predictors found in the current study to contribute to
disordered eating (family behavior control and maternal acceptance-rejection), Bruch
(1973) argued that anorexia may be interpreted as a result of a child’s struggle to develop
her own self-identity within a mother-daughter relationship that limits the daughter’s
autonomy. This limited autonomy may be related to the perception of high behavior
control within the family as well as increased maternal rejection. Therapeutically it is
often desired to determine the functionality of the primary symptom. According to
Bruch’s (1973) contention, combined with the ﬁndings of this study, the purpose of an
eating disorder may be to break free from her mother’s prescribed identity and develop
her own while gaining control in some area of her life. If this hypothesis were captured
by the family therapist, he or she would work to change the family system so that the
young woman would have the freedom to develop her own identity without the need for

an eating disorder to facilitate this yearning. Simultaneously, the family therapist would

125

need to balance the system by concentrating on what function it served for mother to be
critical, rejecting, or enmeshed with her daughter, likely examining her current spousal
relationship or family of original legacies. A number of writers have talked about the
difﬁculty parents may have letting go of their child during the separation-individual
phase of the life cycle, proposing that this struggle may inhibit the growth of the
developing adult.

Beyond considering the most appropriate treatment format for the individual,
conjoint versus separated, it is prudent to acknowledge the experience of the parents in
the family system as being impacted by the presence of an eating disorder. A great
majority of studies, including the current one, consider parents and families as potential
contributors to the etiology of an eating disorder. However, it is important to note that
any time a child is diagnosed with an illness, parents are put at psychological risk as they
are faced with the stress of adapting to the illness (Gilbert, Shaw, & Notar, 2000).
LeGrange, Eisler, Dare, and Russell (p. 348, 1992) state, “we do not know what roles
family changes might play in preventing future relapse of the illness”. In essence,
acknowledging that parents can serve as an asset to treatment rather than viewing them as
a necessary liability is likely beneﬁcial to the treatment process and recovery.

Future directions

A logical next step to the current research would include an investigation looking
at the children’s role as provokers of their own rejection (Rohner, 1991) or the presence
of behavior control, problem-solving, or maternal acceptance-rejection as a response to
disordered eating symptomatology rather than as precursors. Any linear relationship

gathering information from one individual in a system is inherently limited. Therefore

126

future research could strive to include parents and siblings in data reporting to gain a
more holistic picture of the system. As Bruch contended, there is likely a mutual
reciprocation of rejection, parents to child and child to parents. Gaining the parents’
persepectives would add richness and depth to an analysis such as this.

It is important to note the presence of inconsistent conceptualizations of such
constructs as “rejection” and “acceptance” in the literature. Due to the limited research
conducted with disordered eating with the incorporation of “rejection” as a construct, it is
difﬁcult to discuss the similarities and differences in the numerous factors that have been
found in families relating to disordered eating. The Parental Bonding Instrument (FBI) is
a good example of an instrument that may be measuring much the same construct as
parental acceptance and rejection, but has used language mirroring attachment theory.
Because of the differing measuring techniques, great variability exists across studies so
that it becomes difﬁcult to equate what is being measured with certainty or validity.

Greater effort needs to be taken to use language that can be transferable between
studies. Perhaps the next logical step is to create a construct of rejection more broad-
based that includes many of the factors discussed by many of the researchers conducting
studies in the area of disordered eating. This multifactorial construct will likely consist of
a number of variables that have been shown to foster the presence of disordered eatingin
individuals such as family teasing, controlling and critical parenting, a main discourse on
appearance and weight (Haworth-Hoeppner, 2000).

The creation of a rejection construct equipped with both what fosters it and what
results from it would provide clinicians working with eating disordered populations a

variety of entrances into the treatment system. For instance, one option may be to work

127

with the parental unit in the case where perceived rejection is present in order to shift the
discourse in the family from appearance and weight based. Another option may be to
work with the individual who feels rejected on ways to alter this

Further, the academic body of literature would beneﬁt from the inclusion of a
clinical sample to test the hypotheses proposed in the current study. Conﬁrmatory clinical
diagnoses and speciﬁed symptoms (restricting, purging, binge-eating) would certainly
enhance the ability to conclude, with more accuracy, the relative impact of parental

rejection and perception of unhealthy family functioning on the presence of the disorder.

128

APPENDICES

129

APPENDIX A: Informed Consent

 

 

Consent Form for Participation in a Research Study

 

Explanation of Research
The research study you are being asked to participate in is looking at the relationship between

individual’s perception of their family environment and their relationship with food, eating, and their
body. The goal is to solicit information that can be used to better understand the experiences and
needs of young adults. All of the information you provide to the investigators will be used for research

purposes only.

Expectations for Participaiion

If you agree to participate in this research study you will be asked to complete ﬁve questionnaires.
The ﬁve questionnaires will take approximately 1 hour to 1 1/2 hours to complete. The completion of
the study’s questionnaires is the only expectation for participation in the study.

Voluntary NMe

Participation in this research study is purely voluntary. If you decide to participate, you are free to
withdraw at any time without affecting any relationships with Michigan State University. You may
otherwise refuse to answer any questions if you decide to consent to participate in the study. You may
discontinue your participation at any time without penalty, meaning that withdrawal from the research
study will not affect your course grade. Again, your participation is voluntary and it is within your full
discretion to continue or withdraw from the study at any point in time.

Risks and Beneﬁts
There are not expected to be any risks or beneﬁts as a result of your participation in this study.

Conﬁdentiality
All information provided will be kept conﬁdential. No identifying information, such as name or

address, will be requested on the questionnaires. All questionnaires will be protected through storage
in a locked ﬁle cabinet to which only the investigators have access. The only people who will have
access to questionnaire responses and other evidence of participation will be the projects investigators,
Dr. Marsha Carolan and Tianna Hoppe-Rooney. Your privacy will be protected to the maximum
extent allowable by law.

Contacts and Questions

If you have any questions about this study, please contact the primary investigator Tianna Hoppe-
Rooney at (517) 367-5749 or Dr. Marsha Carolan at (517) 432-3327, Department of Family and Child
Ecology, 138 Human Ecology Building, Michigan State University, East Lansing, MI 48824. Tianna
Hoppe-Rooney can be reached by e-mail at hopperoo@msu.edu or Dr. Carolan can also be reached by
e—mail at grolan@m_su.edu.

If you have any questions or concerns regarding your rights as a study participant, or are dissatisﬁed at
any time with any aspect of this study, you may contact — anonymously, if you wish, Ashir Kumar,
M.D., Chair of the University Committee on Research Involving Human Subjects (UCRIHS) by
phone: (517) 355-2180, fax: (517)432-4503, e-mail: ucrihs@msu.edu, or regular mail: 202 Olds Hall,
East Lansing, MI 48824.

A copy of this form will be provided for your records.

 

Statement of Consent

By signing below I agree to participate in the research study being conducted through the Department
of Family and Child Ecology at Michigan State University. Your signature below indicates your
voluntary agreement to participate in this study.

Name: Date:

 

130

 

APPENDIX B: Demographic Questionnaire
Individual Information

1. Sex.
1. Female
2. Male

2. Age.

1. l8-l9
2. 20-21
3. 22-23
4. 24-25
5. 25+

 

3. Education Level. What is the highest grade you completed in school?
1. Freshman

2. Sophomore

3. Junior

4. Senior

5. Graduate

 

4. Major. What is your major? (i.e. child development, psychology, business)

 

5. Athletics. Are you part of a collegiate-level sports team?
1. Yes
2. No

6. Greek Afﬁliation. Are you part of the Greek system within MSU (sorority or
fraternity)
1. Yes
2. No

7. Employment. Are you currently employed?

1. Unemployed; not looking for work

2. Unemployed; looking for work

3. Employed part-time (includes summer employment)
4. Employed full-time

5. Other (please specify)

 

 

131

Demographic Questionnaire continued
Family Information

8. Relationship of parents. Check all the following that are true.
1. Married and living together

2. Not married but living together (consensual union)
3. Separated (married, but not living together)

4. Divorced

5. Widowed

6. Never married

 

9. Household income. What is your household’s income level per year?
1. $15,000 or less
__ 2. $15,000 - $35,000
3. $35,000 - $50,000
4. $50,000 - $75,000
5. $75,000 or above

10. Ethnicity. What is your family’s ethnic background? (check all that apply)
1. Caucasian

2. African American

3. Asian American / Paciﬁc Islander

4. Hispanic American

5. Native American

6. Other:

 

 

l 1. Children. How many siblings do you have?

 

12. Birth order. Putting all your siblings in order from oldest (ﬁrst born) to youngest (last
born), where do you fall (e. g., only child, ﬁrst born, second born, etc.)?

 

132

APPENDIX C: Eating Attitudes Test (EAT-26)

Please check a response for each of the following questions.

 

Always Usually Often Sometimes Rarely Never

 

1. Am terriﬁed about being overweight.

 

2. Avoid eatingwhen I am hungry.

 

3. Find myself preoccupied with food.

 

4. Have gone on eating binges where I feel that

I may not be able to stop.

 

5. Cut my food into small pieces.

 

6. Aware of the calorie content of foods that I eat.

 

7. Particularly avoid food with a high carbohydrate
content (i.e., bread, rice, potatoes, etc.).

 

8. Feel that others would Jrefer if I ate more.

 

9. Vomit after I have eaten.

 

10. Feel extremely guilty after eating.

 

11. Am preoccupied with a desire to be thinner.

 

12. Think about burning up calories when I exercise.

 

13. Other people think that I am too thin.

 

14. Am preoccupied with the thought of having
fat on my body.

 

15. Take longer than others to eat my meals.

 

16. Avoid foods with sugar in them.

 

17. Eat diet foods.

 

18. Feel that food controls my life.

 

19. Display self-control around food.

 

20. Feel that others pressure me to eat.

 

21. Give too much time and thought to food.

 

22. Feel uncomfortable after eating sweets.

 

23. Engage in dieting behavior.

 

24. Like my stomach to be empty.

 

25. Enjoy trying new rich foods.

 

 

26. Have the impulse to vomit after meals.

 

Copyright: Garner, Olmsted, Bohr, and Garﬁnkel, 1982

133

 

APPENDIX D: Parental Acceptance Rejection Questionnaire (Mother)

 

 

Name Date

The following pages contain a number of statements describing the way different mothers
act toward their children. Read each statement carefully and think how well it describes
the way your biological mother treated you while you were growing up. Especially think
about the time when you were about 7-12 years old. Work quickly; give your ﬁrst
impression and move on to the next item. Do not dwell on any item.

Four lines are drawn after each sentence. If the statement is basically true about the way
your mother treated you then ask yourself, “Was it almost always true?” or “Was it only
sometimes true?” If you think your mother almost always treated you that way, put an X
on the line ALMOST ALWAYS TRUE; if the statement was sometimes true about the
way your mother treated you then mark SOMETIMES TRUE. If you feel the statement is
basically untrue about the way your mother treated you then ask yourself, “Was it rarely
true?” or “Was it almost never true?” If it is rarely true about the way your mother treated
you put an X on the line RARELY TRUE; if you feel the statement is almost never true
then mark ALMOST NEVER TRUE.

Remember, there is no right or wrong answer to any statement, so be as honest as you
can. Answer each statement the way you feel your mother really is rather than the way
you might like her to be. For example, if in your memory your mother almost always
hugged and kissed you when you were good you should mark the item as follows:

EXAMPLE
TRUE OF NOT TRUE
MY MOTHER OF MY MOTHER
Almost Almost
Always Sometimes Rarely Never
True True Tme True

1. My mother hugged and kissed me
when I was good... X

Copyright Ronald P. Rohner, 1976

134

MY MOTHER

I. Said nice things about me...

2. Nagged or scolded me when I am bad...

3. Totally ignored me...
4. Did not really love me...

5. Talked to me about our plans and
listened to what I had to say...

6. Complained about me to others
when I did not listen to her...

7. Took an active interest in me...

8. Encouraged me to bring my friends
home, and tried to make things
pleasant for them...

9. Ridiculed and made fun of me...

10. Ignored me as long as I did not do
anything to bother her...

11. Yelled at me when she was angry...

12. Made it easy for me to tell her
things that were important...

13. Treated me harshly...

l4. Enjoyed having me around her...

15. Made me feel proud when I did well...

16. Hit me, even when I did not deserve it...

TRUE OF

135

MY MOTHER
Almost
Always Sometimes
True True

NOT TRUE

Rarely
True

OF MY MOTHER

Almost
Never
True

TRUE OF NOT TRUE

 

MY MOTHE OF MY MOTHER
Almost Almost
Always Sometimes Rarely Never
True True True True
MY MOTHER
l7. Forgot things she was supposed to
do for me...

18. Saw me as a big bother/sister...
l9. Praised me to others...

20. Punished me severely when she was angry.

21. Made sure I had the right kind of
food to eat...

22. Talked to me in a warm and loving way...
23. Got angry at me easily...

24. Was too busy to answer my questions...
25. Seemed to dislike me...

26. Said nice things to me when I deserved
them...

27. Got mad quickly and picked on me...
28. Was concerned who my friends were...
29. Was really interested in what I did...
30. Said many unkind things to me...

31. Ignored me when I asked for help...

32. Thought it is my own fault when I was
having trouble...

33. Made me feel wanted and needed...

34. Told me that I got on her nerves...

136

MY MOTHER

35. Paid a lot of attention to me...

36. Told me how proud she was of me
when I was good...

37. Went out of her way to hurt my
feelings...

38. Forgot important things I thought
she should remember...

39. Made me feel I was not loved any
more if I misbehaved. ..

40. Made me feel what I did was important...

41. Frightened or threatened me when I did
something wrong...

42. Liked to spend time with me...

43. Tried to help me when I was scared
or upset...

44. Shamed me in front of my friends
when I misbehaved. ..

45. Tried to stay away from me...
46. Complained about me...

47. Cared about what I thought and liked me
to talk about it...

48. Felt other children were better than I
was no matter what I did...

49. Cared about what I would like when
she made plans...

TRUE OF

MY MOTHER
Almost
Always Sometimes
True True

137

NOT TRUE

OF MY MOTHER

Rarely
True

Almost
Never
True

 

 

 

 

 

TRUE OF NOT TRUE

MY MOTHER OF MY MOTHER
Almost Almost
Always Sometimes Rarely Never
True True True True

MY MOTHER

50. Let me do things I thought were important,
even if it was inconvenient for her...

51. Thought other children behaved better
than I did...

52. Made other people take care of me
(for example, a neighbor or relative).

53. Let me know I was not wanted...
54. Was interested in the things I did...

55. Tried to make me feel better when I
was hurt or sick...

56. Told me how ashamed she was when
I misbehaved. ..

57. Let me know she loved me.
58. Treated me gently and with kindness...

59. Made me feel ashamed or guilty
when I misbehaved. ..

60. Tried to make me happy...

138

APPENDD( E: Parental Acceptance Rejection Questionnaire (Father)

 

 

Name Date

The following pages contain a number of statements describing the way different fathers
act toward their children. Read each statement carefully and think how well it describes
the way your biological father treated you while you were growing up. Especially think
about the time when you were about 7-12 years old. Work quickly; give your ﬁrst
impression and move on to the next item. Do not dwell on any item.

Four lines are drawn after each sentence. If the statement is basically true about the way
your father treated you then ask yourself, “Was it almost always true?” or “Was it only
sometimes true?” If you think your father almost always treated you that way, put an X
on the line ALMOST ALWAYS TRUE; if the statement was sometimes true about the
way your father treated you then mark SOMETIMES TRUE. If you feel the statement
was basically untrue about the way your father treated you then ask yourself, “Was it
rarely true?” or “Was it almost never true?” If it was rarely true about the way your
father treated you put an X on the line RARELY TRUE; if you feel the statement was
almost never true then mark ALMOST NEVER TRUE.

Remember, there is no right or wrong answer to any statement, so be as honest as you
can. Answer each statement the way you feel your mother really is rather than the way
you might like her to be. For example, if in your memory your father almost always
hugged and kissed you when you were good you should mark the item as follows:

EXAMPLE
TRUE OF NOT TRUE
MY FATHER OF MY FATHER
Almost Almost
Always Sometimes Rarely Never
True True True True
1. My father hugged and kissed me
when I was good... , _X_

Adapted from : Ronald P. Rohner, 1976

139

TRUE OF NOT TRUE

MY FATHER OF MY FATHER
Almost Almost
Always Sometimes Rarely Never
True True True True

MY FATHER

l. Said nice things about me...

2. Nagged or scolded me when I am bad...
3. Totally ignored me...

4. Did not really love me...

5. Talked to me about our plans and
listened to what I had to say...

6. Complained about me to others
when I did not listen to him...

7. Took an active interest in me...

8. Encouraged me to bring my friends
home, and tried to make things
pleasant for them...

9. Ridiculed and made fun of me...

10. Ignored me as long as I did not do
anything to bother him...

11. Yelled at me when he was angry...

12. Made it easy for me to tell him
things that were important...

13. Treated me harshly...

l4. Enjoyed having me around him...

15. Made me feel proud when I did well...
16. Hit me, even when I did not deserve it...

17. Forgot things he was supposed to
do for me...

140

TRUE OF NOT TRUE

MY FATHER OF MY FATHER
Almost Almost
Always Sometimes Rarely Never
True True True True

MY FATHER

18. Saw me as a big bother/sister...
l9. Praised me to others...

20. Punished me severely when he was angry..

21. Made sure I had the right kind of
food to eat...

22. Talked to me in a warm and loving way...
23. Got angry at me easily...

24. Was too busy to answer my questions...
25. Seemed to dislike me...

26. Said nice things to me when I deserved
them...

27. Got mad quickly and picked on me...
28. Was concerned who my friends were...
29. Was really interested in what I did...
30. Said many unkind things to me...

31. Ignored me when I asked for help...

32. Thought it is my own fault when I was
having trouble. . .

33. Made me feel wanted and needed...
34. Told me that I got on his nerves...

35. Paid a lot of attention to me...

141

MY FATHER

36. Told me how proud he was of me
when I was good...

37. Went out of his way to hurt my
feelings...

38. Forgot important things I thought
he should remember...

39. Made me feel I was not loved any
more if I misbehaved. ..

40. Made me feel what I did was important...

41. Frightened or threatened me when I did
something wrong...

42. Liked to spend time with me...

43. Tried to help me when I was scared
or upset...

44. Shamed me in front of my friends
when I misbehaved. ..

45. Tried to stay away from me...
46. Complained about me...

47. Cared about what I thought and liked me
to talk about it...

48. Felt other children were better than I
was no matter what I did...

49. Cared about what I would like when
he made plans...

TRUE OF

MY FATHER
Almost
Always Sometimes
True True

 

50. Let me do things I thought were important,

even if it was inconvenient for him...

142

NOT TRUE
OF MY FATHER

Rarely
True

Almost
Never
True

TRUE OF NOT TRUE

 

MY FATHER OF MY FATHER
Almost Almost
Always Sometimes Rarely Never
True True True True
MY FATHER
51. Thought other children behaved better
than I did...

52. Made other people take care of me
(for example, a neighbor or relative).

53. Let me know I was not wanted...
54. Was interested in the things I did...

55. Tried to make me feel better when I
was hurt or sick...

56. Told me how ashamed he was when
I misbehaved. ..

57. Let me know he loved me...
58. Treated me gently and with kindness...

59. Made me feel ashamed or guilty
when I misbehaved. ..

60. Tried to make me happy...

143

APPENDIX F: Family Assessment Device (FAD)
Instructions:

This booklet contains a number of statements about families. Please read each statement
carefully, and decide how well it describes your own family. You should answer
according to how you see your family.

For each statement there are four (4) possible responses:

Strongly Agree (SA) Check SA if you feel that the statement
describes your farme very accurately

Agree (A) Check A if you feel that the statement
describes your family for the most part.

Disagree (D) Check D if you feel that the statement does
not describe your family for the most part.

Strongly Disagree (SD) Check SD if you feel that the statement does
not describe your family at all.
These four responses will appear below each statement like this:
41. We are not satisﬁed with anything short of perfection

SA A D SD

 

The answer spaces for statement 41 would look like this. For each statement in
the booklet, there is an answer space below. Do not pay attention to the blanks at the far
right-hand side of each space. They are for ofﬁce use only.

Try not to spend too much time thinking about each statement, but respond as
quickly and as honestly as you can. If you have trouble with one, answer with your ﬁrst
reaction. Please be sure to answer m statement and mark all your answers in the spag
provided below each statement.

144

_F_amily Assessment Device

l.

10.

ll.

12.

l3.

14.

15.

SA A D SD

SA A D SD

 

 

 

 

 

Planning family activities is difﬁcult because we misunderstand each other.
SA A D SD

 

 

We resolve most everyday problems around the house.
SA A D SD

 

When someone is upset the others know why.
SA A D SD

 

 

When you ask someone to do something, you have to check that they did it.
SA A D SD

 

 

If someone is in trouble, the others become too involved.
SA A D SD

 

In times of crisis we can turn to each other for support.
SA A D SD

 

 

We don’t know what to do when an emergency comes up.
SA A D SD

 

 

We sometimes run out of things we need.

 

We are reluctant to show our affection for each other.
SA A D SD

 

We make sure members meet their family responsibilities.

 

We cannot talk to each other about the sadness we feel.
SA A D SD

 

We usually act on our own decisions regarding problems.
SA A D SD

 

You only get the interest of others when something is important to them.
SA A D SD

 

You can’t tell how a person is feeling from what they are saying.
SA A D SD

 

Family tasks don’t get spread around enough.
SA A D SD

 

145

Famin Assessment Device

l6.

l7.

l8.

19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

29.

30.

Individuals are accepted for what they are.
SA A D SD

 

 

You can easily get away with breaking the rules.
SA A D SD

 

 

People come right out and say things instead of hinting at them.
SA A D SD

 

Some of us just don’t respond motionally.
SA A D SD

 

We know what to do in an emergency.
SA A D SD

 

 

We avoid discussing our fears and concerns.
SA A D SD

 

 

It is difﬁcult to talk to each other about tender feelings.
SA A D SD

 

We are having trouble meeting our bills.
SA A D SD

 

 

After our family tries to solve a problem, we usually discuss whether it worked or
not.
SA A D SD

 

 

We are too self—centered.
SA A D SD

 

 

We can express feelings to each other.
SA A D SD

 

 

We have no clear expectations about toilet habits.
SA A D SD

 

We do not show our love for each other.
SA A D SD

 

 

We talk to people directly rather than through go-betweens.
SA A D SD

 

 

Each of us has particular duties and responsibilities.
SA A D SD

 

146

_F_amily Assessment Device

31.

32.

33.

34.

35.

36.

37.

38.

39.

40.

41.

42.

43.

45.

There are lots of bad feelings in the family.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

SA A D SD
We have rules about hitting people.
SA A D SD
We get involved with each other only when something interests us.
SA A D SD
There’s little time to explore personal interests.
SA A D SD
We often don’t say what we mean.
SA A D SD
We feel accepted for what we are.
SA A D SD
We show interest in each other when we can get something out of it personally.
SA A D SD
We resolve most emotional upsets that come up.
SA A D SD
Tenderness takes second place to other things in our family.
SA A D SD
We discuss who is to do household jobs.
SA A D SD
Making decisions is a problem for our family.
SA A D SD
Our family shows interest in each other only when they can get something out of it.
SA A D SD
We are frank with each other.
SA A D SD

 

 

 

We don’t hold to any rules or standards.
SA A D SD

 

 

If people are asked to do something, they need reminding.
SA A D SD

 

 

147

Family Assessment Device

46.

47.

48.

49.

50.

51.

52.

53.

54.

55.

56.

57.

58.

59.

60.

We are able to make decisions about how to solve problems.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

SA A D SD
If the rules are broken, we don’t know what to expect.
SA A D SD
Anything goes in our family.
SA A D SD
We express tenderness.
SA A D SD
We confront problems involving feelings.
SA A D SD
We don’t get along well together.
SA A D SD
We don’t talk to each other when we are angry.
SA A D SD
We are generally dissatisﬁed with the family duties assigned to us.
SA A D SD
Even though we mean well, we intrude too much into each others lives.
SA A D SD
There are rules about dangerous decisions.
SA A D SD
We conﬁde in each other.
SA A D SD
We cry openly.
SA A D SD

 

 

We don’t have reasonable transport.
SA A D SD

 

 

When we don’t like what someone has done, we tell them.
SA A D SD

 

 

We try to think of different ways to solve problems.
SA A D SD

 

 

148

REFERENCES

American Psychiatric Association (1994). Diagnostic and statistical manual of mental
disorders. Washington, DC: APA.

Barnow, S., Schuckit, M., Lucht, M., John, U., & Freyberger, H. (2002). The importance
of a positive family history of alcoholism, parental rejection and emotional

warmth, behavioral problems and peer substance use for alcohol problems in
teenagers: A path analysis. Journal of Studies on Alcohol, 63, 305-315.

Bowlby, J. (1966). Maternal care and mental health. New York, NY: Schocken.

Bronfenbrenner, U. (1979). The ecology of human development: Experiments by nature
and design. Cambridge, MA: Harvard University Press.

Bronfenbrenner, U. (1995). Developmental ecology through space and time: A future
perspective. In P. Moen, G. Elder Jr., and K. Luscher (Eds.), Examining lives in
context, (pp. 619-647). Washington, DC: American Psychological Association.

Bronfenbrenner, U. (1999). Environments in developmental perspective: Theoretical and
operational models. In S.L. Friedman and TD. Wachs (Eds.), Measuring
environment across the life span, (pp. 328). Washington, DC: American
Psychological Association.

Bronfenbrenner, U. & Ceci, S. (1994). N ature-nurture reconceptualized in developmental
perspective: A bioecological model. Psychological Review, 101(4), 568-586.

Bruch, H. (1973). Eating disorders: Obesity, anorexia nervosa and the person within.
New York: Basic Books.

Bruch, H. (1974). Eating disorders: Obesity, anorexia nervosa and the person within.
London: Routledge & Kegan Paul.

Bubolz, M. & Sontag, MS. (1993). Human ecology theory. In P. Boss, W. Doherty, R.
W. Schumm, & S. Steinmetz (Eds.), Sourcebook of family theories and methods:
A contextual approach (pp. 419-448). New York, NY: Plenum Publishing.

Bubolz, M., Eicher, 1., & Sontag, MS. (1979). The human ecosystem: A model. Journal
of Home Economics, 71, 28-31.

Buchner, A., Erdfelder, E., & Paul, F. (1997). How to Use G*Power [WWW document].

URL http://www.psycho.uni-
duesseldorf.de/aap/projects/gpower/how_to_use_gpower.htrnl

149

 

Burgess, E.W. (1926). The family as a unity of interacting personalities. The Family, 7, l.-
3-9.

Burr, W., Leigh, G., Day, R., & Constantine, J. (1979). Symbolic interaction and the
family. In W. Burr, R. Hill, FL Ney, and I. Reiss (Eds.), Contemporary theories
about the family, (pp. 42-108). New York, NY: The Free Press.

Campo, A. & Rohner, R. (1992). Relationships between perceived parental acceptance-
rejection, psychological adjustment, and substance abuse among youth adults.
Child Abuse and Neglect, 16, 429-440.

Casper, R. & Troiani, R. (2001). Family functioning in anorexia nervosa differs by
subtype. International Journal of Eating Disorders, 30, 339-342.

Cast, A. (2003). Power and the ability to deﬁne the situation. Social Psychology
Quarterly, 66(3), 185-201.

Cohen, J. (1977). Statistical power analysis for the behavioral sciences. New York, NY:
Academic Press.

Crowther, J.H., Wolf, E.M., & Sherwood, N. (1992). Epidemiology of bulimia nervosa.
In M. Crowther, D.L. Tennenbaum. S.E. Hobfoll, and M.A.P. Stephens (Eds.),
The etiology of bulimia nervosa: The individual and familial context, (pp. 1-26).
Washington, DC: Taylor & Francis.

Dare, C., Eisler, 1., Russell, 6., & Szmukler, G. (1990). The clinical and theoretical
impact of a controlled trial of family therapy in anorexia nervosa. Journal of
Marital & Family Therapy, 16, 39-57.

Dolan, B., Lieberman, 8., Evans, C., & Lacey, J. (1990). Family features associated with
normal body weight bulimia. International Journal of Eating Disorders, 9(6),
639-647.

Dorrriny, N., Johnson, W.B., & Koch, C. (2000). Perceptionof parental acceptance in
women with binge eating disorder. The Journal of Psychology, 134, 23-36.

Eisler, 1., Dare, C., Russell, G., Szmukler, G., LeGrange, D., & Dodge, E. (1997). Family
and individual therapy in anorexia nervosa. A 5-year follow-up. Archives of
General Psychiatry, 54, 1025-1030.

Envoy, J. (1981). The Rejected: Psychological consequences of parental rejection.
University Park, PA: The Pennsylvania State University Press.

Epstein, N. Baldwin, L. & Bishop, D. (1983). The McMaster Family Assessment Device.
Journal of Marital and Family Therapy, 9(2), 171-180.

150

 

Epstein, N ., Bishop, D., Levin, S. (1978). The McMaster model of family functioning.
Journal of Marriage and Family Counseling, 19—3 1.

Epstein, N., Bishop, D., Ryan, C., Miller, I., & Keitner, G. (1993). The McMaster model:
View of healthy family functioning. In F. Walsh (Ed.), Normal Family Processes,
Second Edition, (pp. 138-160). New York, NY: Guilford Press. ‘

Fairbum, C., & Beglin, S. (1990). Studies of the epidemiology of bulimia nervosa.
American Journal of Psychiatry, 147, 401-408.

Fairbum, C., Hay, P., & Welch, S. (1993). Binge eating and bulimia nervosa:
Distribution and determinants. In C.G. Fairbum & G.T. Wilson, (Eds.), Binge
eating: Nature, assessment, and treatment, (pp. 123-143). New York, NY:
Guilford.

Fairbum, C., Welch, S., & Doll, H. (1997). Risk factors for bulimia nervosa: A
community-based case-control study. Archives of General Psychiatry, 54, 509-
517.

Felker, K. & Stivers, C. (1994). The relationship of gender and family environment to
eating disorder risk in adolescence. Adolescence, 29(116), 821-834.

Fenichel, E. (2002). Relationships at risk: The policy environment as a context for infant
development. Infants and Young Children, 15(2), 49-56.

Fonseca, H., Ireland, M., & Resnick, M. (2002). Familial correlates of extreme weight

control behaviors among adolescents. International Journal of Eating Disorders,
32, 441-448.

Freedman Melson, G. (1980). Family and environment. Minneapolis, MN: Burgess
Publishing Company.

Friedman, L. (1997). The role of the father-daughter relationship in eating disorders. A

dissertation. Widener University, School of Human Service Professions; Chester,
PA.

Garner, D. & Garﬁnkel, P. (1979). The eating attitudes test: An index of the symptoms of
anorexia nervosa, Psychological Medicine, 9, 273-279.

Garner, D., Olmsted, M., Bohr, Y., & Garﬁnkel, P. (1982). The eating attitudes test:

Psychometric features and clinical correlates. Psychological Medicine, 12, 871-
878.

Gilbert, A., Shaw, 3., & Notar, M. (2000). The impact of eating disorders on family
relationships. Eating Disorders, 8, 331-345.

151

Griffore, R. & Phenice, L. (2003). The language of human ecology: A general systems
perspective. Dubuque, IA: Kendall/Hunt Publishing Company.

Grotevant, H. & Carlson, C. (1989). Family assessment: A guide to methods and
measures. New York, NY: Guilford Press.

Gutzwiller-Jurman, K. (1999). Relationship between eating disorder symptomatology and
perceptions of closeness-caregiving and intrusiveness in the mother-daughter and
father-daughter relationship. A dissertation. San Diego: California School of
Professional Psychology.

Halmi, K., Frank, J. & Schwartz, E. (1981). Binge-eating and vomiting: A survey of a
college population. Psychological Medicine, 11, 967-706.

Hawkes, G. (1957). The child in the family. Marriage and Family Living, 19, 46-51.

Haworth-Hoeppner, S. (2000). The critical shapes of body image: The role of culture and
family in the production of eating disorders. Journal of Marriage and the Family,
62, 212-227.

Heatherton, T., Mahamedi, F., Striepe, M., Field, A., & Keel, P. (1997). A 10-year
longitudinal study of body weight, dieting, and eating disorder symptoms. Journal
of Abnormal Psychology, 106, 117-125.

Hesse-Biber, S., Marino, M., Watts-Roy, D. (1999). A longitudinal study of eating
disorders among college women. Gender & Society, 13(3), 385-408.

Hodges, E., Cochrange, C., & Brewerton, T. (1998). Family characteristics of binge-
eating disorder patients. International Journal of Eating Disorders, 7, 51-61.

Hoffman, L. (1990). Constructing realities: An art of lenses. Family Process, 29(1), 1-12.

Humphrey, L. (1986). Structural analysis of parent-child relationships in eating disorders.
Journal of Abnormal Psychology, 95(4), 395-402.

Humphrey, L. (1988). Relationships within subtypes of anorexic, bulimic, and normal
families. Journal of the American Academy of Child and Adolescent Psychiatry,
27, 544-551.

Humphrey, L. (1989). Observed family interactions among subtypes of eating disorders
using structural analysis of behavior. Journal of Consulting and Clinical
Psychology, 57, 206-214.

Ingoldsby, B., Smith, S., & Miller, J.E. (2004). Exploring family theories. Symbolic

Interactionism Theory, Chapter 4, pp. 81-92. Los Angeles, CA: Roxbury
Publishing Company.

152

Johnson, C., & Flach, A. (1985). Family characteristics of 105 patients with bulimia.
American Journal of Psychiatry, 142, 1321-1324.

Kabacoff, R., Miller, 1., Bishop, D., Epstein, N., & Keitner, G. (1990). A psychometric
study of the McMaster Family Assessment Device. Journal of Family
Psychology, 3, 431-439.

Kanakis, D. & Thelen, M. (1995). Parental variables associated with bulimia nervosa.
Addictive Behaviors, 20(4), 491-500.

Kent, J. & Clopton, J. (1988). Bulimia: A comparison of psychological adjustment and
familial characteristics in a nonclinical sample. Journal of Clinical Psychology,
44, 964-971.

Khaleque, A. & Rohner, R. (2002). Perceived parental acceptance-rejection and
psychological adjustment: A meta-analysis of cross-cultural and intracultural
studies. Journal of Marriage and Family, 64, 54-64.

Kichler, J. C. & Crowther, J. H. (2001). The effects of maternal modeling and negative
familial communication on women’s eating attitudes and body image. Behavior
Therapy, 32, 443—457.

Kog, E. & Vandereycken, W. (1989). Family interaction in eating disorder patients and
normal controls. International Journal of Eating Disorders, 8, 11-23.

Koszewski, W., Newell, B., & Higgins, J. (1990). Effect of a nutrition education program
on the eating attitudes and behaviors of college women. Journal of College
Student Development, 31, 203-210.

LaRossa, R. & Reitzes, D. (1993). Symbolic interactionism and family studies. In P.
Boss, W. Doherty, R. LaRossa, W. Schumm, and S. Steinmetz (Eds.), Sourcebook
of Family Theories and Methods: A Contextual Approach, (pp. 135-163). New
York, NY: Plenum Press.

Latzer, Y., Hochdorf, Z., Bachar, E., & Canetti, L. (2002). Attachment style and family
functioning as discriminating factors in eating disorders. Contemporary Family
Therapy, 24, 581-599).

LeGrange, D., Eisler, 1., Dare, C., & Russell, G. (1992). Evaluation of family treatments
in adolescent anorexia nervosa: A pilot study. International Journal of Eating
Disorders, 12, 347-357.

Lewis-Beck, M. (1980). Applied Regression. Newbury Park, CA: Sage Publications.

Longmore, M. (1998). Symbolic interactionism and the study of sexuality. The Journal of
Sex Research, 35(1), 44-57.

153

MacBrayer, E., Smith, G., McCarthy, D., Demos, S. & Simmons, J. (2001). The role of

family of origin food-related experiences in bulimic symptomatology.
International Journal of Eating Disorders, 30, 149-160.

Maine, M. (1991). Father hunger: Fathers, daughters, and food. Carlsbad, CA: Gurze
Books.

McCauley Ohannessian, C., Lerner, R., von Eye, A., & Lerner, J. (1996). Direct and
indirect relations between perceived parental acceptance, perceptions of the self,

and emotional adjustment during early adolescence. Family and Consumer
Sciences Research Journal, 25(2), 159-183.

McClendon, M. (1994). Multiple regression and causal analysis. Prospect Heights,
IL: Waveland Press, Inc.

McGrane, D. & Carr, A. (2002). Young women at risk for eating disorders: Perceived
family dysfunction and parental psychological problems. Contemporary Family
Therapy, 24(2), 385-395.

McKinley, N .M. (1999). Women and objectiﬁed body consciousness: Mothers’ and
daughters’ body experience in cultural, developmental, and familial context.
Developmental Psychology, 35(3), 760-769.

Mead, G. (1934/ 1956). On Social Psychology: Selected Papers. (Anselem Strauss, Ed.).
Chicago, IL: University of Chicago Press.

Miller, I., Epstein, N., Bishop, D., & Keitner, G. (1985). The McMaster Family
Assessment Device: Reliability and validity. Journal of Marital and Family
Therapy, 11(4), 345-356.

Mintz, L. & Betz, N. (1988). Prevalence and correlates of eating disordered behaviors
among undergraduate women. Journal of Counseling Psychology, 35, 463-471.

Minuchin, S. & Fishman, H. (1979). The psychosomatic family in child psychiatry.
Journal of the American Academy of Child Psychiatry, 18, 76-84.

Minuchin, S., Rosman, B.L., and Baker, L. (1978). Psychosomatic Families: Anorexia
Nervosa in Context. Cambridge, MA: Harvard University Press.

Moreno, A. & Thelen, M. (1993). Parental factors related to bulimia nervosa. Addictive
Behaviors, 18, 681-689.

Murdock, M. (1992). Daughters in the father’s world. In C.Scull (Ed.), Fathers, sons, &

daughters. Exploring fatherhood, renewing the bond, (pp. 107-114). Los Angeles:
Jeremy P. Tarcher, Inc.

154

Mussel] Pederson, M., Binford, R. & Fulkerson, J. (2000). Eating disorders: Summary of
risk factors, prevention programming, and prevention research. Counseling
Psychologist 28(6), 764-796.

Nelson, W., Hughes, H., Katz, B., & Searight, HR. (1999). Anorexic eating attitudes and
behaviors of male and female college students. Adolescence, 34(135), 621-633.

Netemeyer, S. & Williamson, D. (2001). Assessment of eating disturbance in children
and adolescents with eating disorders and obesity. In J .K. Thompson and L.
Smolak (Eds.), Body Image, Eating Disorders and Obesity in Youth, (pp. 215-
236). Washington, DC: American Psychological Association..

Neumark-Sztainer, D., Story, M., Harman, P., Beuhring, T., & Resnick, M. (2000).
Disordered eating among adolescents: Associations with sexual/physical abuse

and other familial/psychosocial factors. International Journal of Eating
Disorders 28, 249-258.

Ordman, A. & Kirschenbaum, D. (1986). Bulimia: Assessment of eating, psychological
adjustment, and familial characteristics. International Journal of Eating
Disorders, 5, 865-878.

Phares, V. (1992). Where’s poppa? The relative lack of attention to the role of fathers in
child and adolescent psychological. American Psychologist, 47, 656-664.

Pope, H., Hudson, J ., Yurglen-Todd, D., & Hudson, M. (1984). Prevalence of anorexia
and bulimia in three student populations. International Journal of Eating
Disorders, 3, 2-52.

Rohner, R. (1986). The warmth dimension: Foundations of parental acceptance-rejection
theory. Beverly Hills, CA: Sage Publications.

Rohner, R. (1991). Handbook for the study of parental acceptance and rejection (rev.
ed.). Storrs, CT: The University of Connecticut, Center for the Study of Parental
Acceptance and Rejection.

Rohner, R. (2002). Acceptance-rejection bibliography. Center for the Study of Parental
Acceptance and Rejection. Retrieved from httpzllvm.uconn.edu/~rohner

Rohner, R‘. & Britner, P. (2002). Worldwide mental health correlates of parental
acceptance-rejection: Review of cross-cultural and intracultural evidence. Cross-
Cultural Research, 36(1), 16-47.

Rohner, R. & Nielsen, C. (1978). Parental acceptance and rejection: A review and
annotated bibliography of research and theory, vol. I. New Haven, CT: Human
Relations Area Files.

155

Rothbaum, F. & Weisz, J. (1994). Parental caregiving and child externalizing behavior in
non-clinical samples: A meta-analysis. Psychological Bulletin, 116, 55-74.

Rupp, G. & Jurkovic, G. (1996). Familial and individual perspective-taking processes in
adolescent females with bulimic symptomatology. The American Journal of
Family Therapy, 24(1), 75-82.

Russell, G., Szmukler, G., Dare, C., & Eisler, I (1987). An evauation of family therapy in
anorexia nervosa and bulimia nervosa. Archives in General Psychiatry, 44, 1047-
1056.

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

Sackin, H.D. (1985). The parents of children with psychosomatic diseases: A critical
review of the literature. In El. Anthony and G.H. Pollock (Eds.), Parental
Inﬂuences in Health and Disease, (pp. 403-425). Boston, MA: Little, Brown and
Company.

Scalf-Mclver, L. & Thompson, K. (1989). Family correlates of bulimic characteristics in
college females. Journal of Clinical Psychology, 45(3), 467-472.

Scholte, R., van Lieshout, C., & van Aken, M. (2001). Perceived relational support in
adolescence: Dimensions, conﬁgurations, and adolescent adjustment. Journal of

Research on Adolescence, 1 1(1), 71-94.

Schroeder, L., Sjoquist, D., & Stephan, P. (1986). Understanding regression analysis.
Newbury Park, CA: Sage Publications.

Schwitzer, A., Bergholz, K., Dore, T., & Salimi, L. (1998). Eating disorders among
college women: Prevention, education, and treatment responses. College Health,
46, 199-207.

Secunda, V. (1992). Women and their fathers. New York: Delacorte Press.

Selvini Palazzoli, M. (1978). Self-Starvation: From Individual to Family Therapy in the
Treatment of Anorexia Nervosa. London: Chaucer Publishing Co.

Shisslak, C., Crago, M., Estes, L. (1995). The spectrum of eating disturbances.
International Journal of Eating Disorders, 18, 209-219.

Smith, M. & Thelen, M. (1984). Development and validation of a test of bulimia. Journal
of Consulting and Clinical Psychology, 152, 863-872.

156

Steiger, H., Liquornik, K. Chapman, J. & Hussein, N. (1991). Personality and family
disturbances in eating disorder patients; Comparison of “restricters” and

“purgers” to normal controls. International Journal of Eating Disorders, 10, 501-
512.

Steiger, H., Stotland, S, Trottier, J. & Ghadirian, A. (1996). Familial eating concerns and
psychopathological traits: Causal implications of transgenerational effects.
International Journal of Eating Disorders, 19, 147-157.

Steinberg, A. & Phares, V. (2001). Family functioning, body image, and eating
disturbances. In K. Thompson and L. Smolak (Eds.), Body Image, Eating
Disorders, and Obesity in Youth: Assessment, Prevention, and Treatment, (pp.
127-147). Washington, DC: American Psychological Association.

Steinhausen, H. & Winkler Metzke, C. (2001). Risk, compensatory, vulnerability, and

protective factors inﬂuencing mental health in adolescence. Journal of Youth and
Adolescence, 30(3), 259-280.

Stice, E., Kellen, J ., Hayward, C., & Taylor, C. (1998). Support for the continuity
hypothesis of bulimic pathology. Journal of Consulting and Clinical Psychology,
66, 784-790.

Striegel-Moore, R., (1992). Prevention of bulimia nervosa: Questions and challenges. In
J.H. Crowther, D.L. Tennebaum, S.E. Hobfell, and MA. Stephens (Eds.), The
etiology of bulimia nervosa: The individual and familial context. Washington,
DC: Hemisphere.

Striegel-Moore, R., Silberstein, L.R., Frensch, P., & Rodin, J. (1989). A prospective
study of disordered eating among college students. International Journal of
Eating Disorders, 8, 499-509.

Strong, B., DeVault, C., Sayad, B., & Cohen, T. (2001). The marriage and family
experience: Intimate relationships in a changing society. Belmont, CA: Thomson
Learning.

Studart, G., Laraia, M., Ballenger, J., & Lydiard, R. (1990). Early family experiences of
women with bulimia and depression. Archieves of Psychatric Nursing, 4(1), 43-
52.

Thomson Ross, L. & Gill, J. (2002). Eating disorders: Relations with inconsistent

discipline, anxiety, and drinking among college women. Psychological Reports,
91, 289-298.

Twamley, E. & Davis, M. (1999). The sociocultural model of eating disturbance in young

women: The effects of personal attributes and family environment. Journal of
Social and Clinical Psychology, 18(4), 467-489.

157

 

Vandereycken W. & Vanderlinden, J. (1983). Denial of illness and the use of self
reporting measures in anorexia nervosa patients. International Journal of Eating
Disorders, 2, 101-107.

Vanderlinden, J. & Vandereycken, W. (1991). Guidelines for the family therapeutic
approach to eating disorders. Psychotherapy & Psychosomatics, 56, 36-42.

Veneziano, R. (2000). Perceived paternal and maternal acceptance and rural African
American and European American youths’ psychological adjustment. Journal of
Marriage and the Family, 62, 123-132.

Vohs, K., Heatherton, T., & Herrin, M. (2001). Disordered eating and the transition to

college: A prospective study. International Journal of Eating Disorders, 29, 280-
288.

Waller, G. (1994). Bulimic women’s perceptions of interaction within their families.
Psychological Reports, 74, 27-32.

Waller, G., Calam, R., & Slade, P. (1989). Eating disorders and family interaction.
British Journal of Clinical Psychology, 28, 285-296.

Waller, G., Slade, P. & Calam, R. (1990). Family adaptability and cohesion: Relation to
eating attitudes and disorders. International Journal of Eating Disorders, 9(2),
225-228.

Walther, D. (1994). The relationship between perceived family interactional patterns and
symptoms of eating disorders, compared with symptoms of depression. A
dissertation. University of South Carolina, Dept. of Psychology.

Ward, A., Ramsay, R., Tumbull, S., Benedettini, M., & Treasure, J. (2000). Attachment
patterns in eating disorders: Past in the present. International Journal of Eating
Disorders, 28, 370-376.

Webster, J. & Palmer, R. (2000). The childhood and family background of women with
clinical eating disorders: A comparison with women with major depression and
women without psychiatric disorder. Psychological Medicine, 30, 53-60.

Whitbeck, L., Hoyt, D., Simons, R., Conger, R., Elder, G., Lorenz, F., & Huck, S. (1992).
International continuity of parental rejection and depressed affect. Journal of
Personality and Social Psychology, 63(6), 1036-1045.

White, J. (2000). The prevention of eating disorders: A review of the research on risk
factors with implications for practice. Journal of Child and Adolescent
Psychiatric Nursing, 13(2), 76—88.

158

Williamson, D. (1990). Assessment of eating disorders: Obesity, anorexia, and bulimia
nervosa. New York: Pergarnon Press.

Winton, C. (1995). Frameworks for Studying Families. Guilford, CT: The Dushkin
Publishing Group.

Wonderlich, S., Klein, M., & Council, J. (1996). Relationship of social perceptions and

self-concept in bulimia nervosa. Journal of Consulting and Clinical Psychology,
64(6), 1231-1237.

Wonderlich, S. & Swift, W. (1990). Perceptions of parental relationships in the eating
disorders: The relevance of depressed mood. Journal of Abnormal Psychology,
99, 353-360.

Wonderlich, S., Ukestad, L., & Perzacki, R. (1994). Perceptions of non-shared childhood

environment in bulimia nervosa. Journal of the American Academy of Child and
Adolescent Psychiatry, 33, 740-747.

159