University of Southern California Dissertations and Theses

Childhood Sexual Abuse and Depressive Symptoms In A Lesbian Population: An Exploratory Study.

(USC Thesis Other)

Childhood Sexual Abuse and Depressive Symptoms In A Lesbian Population: An Exploratory Study.

PDF

Download

Open document

Flip pages

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content INFORMATION TO USERS
This manuscript has been reproduced from the microfilm master. UMI
films the text directly from the original or copy submitted. Thus, some
thesis and dissertation copies are in typewriter face, while others may be
from any type o f computer printer.
T he quality o f this reproduction is dependent upon the quality o f the
copy submitted. Broken or indistinct print, colored or poor quality
illustrations and photographs, print bleedthrough, substandard margins,
and improper alignment can adversely affect reproduction.
In the unlikely event that the author did not send UMI a complete
manuscript and there are missing pages, these will be noted. Also, if
unauthorized copyright material had to be removed, a note will indicate
the deletion.
Oversize materials (e.g., maps, drawings, charts) are reproduced by
sectioning the original, beginning at the upper left-hand comer and
continuing from left to right in equal sections with small overlaps. Each
original is also photographed in one exposure and is included in reduced
form at the back o f the book.
Photographs included in the original manuscript have been reproduced
xerographically in this copy. Higher quality 6” x 9” black and white
photographic prints are available for any photographs or illustrations
appearing in this copy for an additional charge. Contact UMI directly to
order.
UMI
A Bell & Howell Information Company
300 North Zeeb Road, Ann Arbor MI 48106-1346 USA
313/761-4700 800/321-0600

CHILDHOOD SEXUAL ABUSE AND DEPRESSIVE
SYMPTOMS IN A LESBIAN POPULATION:
AN EXPLORATORY STUDY
by
Christine Diana Cooper
A Dissertation Presented to the
FACULTY OF THE GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(Psychology)
August, 1995
Copyright 1995 Christine Diana Cooper
UMI Number: 9625011
UMI Microform 9625011
Copyright 1996, by UMI Company. All rights reserved.
This microform edition is protected against unauthorized
copying under Title 17, United States Code.
UMI
300 North Zetb Road
Ann Arbor, Ml 48103
UNIVERSITY OF SOUTHERN CALIFORNIA
THE GRADUATE SCHOOL
UNIVERSITY PARK
LOS ANGELES, CALIFORNIA 90007
This dissertation, written by
...............
under the direction of h£C .. Dissertation
Committee, and approved by all its members,
has been presented to and accepted by The
Graduate School, in partial fulfillment of re
quirements for the degree of
DOCTOR OF PHILOSOPHY
Dean of Graduate Studies
Date^±d:...lJ .Lm
DISSERTATION COMMITTEE
Chairperson
Christine Diana Cooper Mitchell Earleywine, PhD
Childhood Sexual Abuse and Depressive Symptoms in a Lesbian
Population: An Exploratory Study
This research was designed to investigate some aspects of depressive symptoms
as they may relate to a reported history of childhood sexual abuse (CSA) in a
population of lesbian women. Survey forms included the Rosenberg Self
Esteem Inventory (Rosenberg-SEI), the Center for Epidemiological Studies-
Depression scale (CES-D), the Parental Bonding Instrument (PBI), and an Early
Sexual Experiences questionnaire. Surveys were distributed at feminist
bookstores, a "coming out day" celebration, and through ads in two
lesbian/feminist newspapers. Forms were completed and returned by 493
lesbian women ranging in age from 16 to 88 (mean age=35.2). Correlational
data demonstrated that, in this sample, lesbian woman who were judged by the
survey criteria to have been sexually abused as children experience higher levels
of depressive symptoms than lesbian women judged not to have experienced
CSA. Path analysis was then used to investigate whether variables derived from
the Rosenberg-SEI and PBI were acting as mediators or moderators o f the
relationship between a reported history of CSA and adult depressive symptoms
as measured by the CES-D. Results were consistent with self-esteem being an
1
important partial mediator of the relationship between a reported history of CSA
and adult depressive symptoms. However, self-esteem was not shown to
moderate the relationship. In addition, results were consistent with the
respondent’s recollection of her mother as "caring" (measured by the PBI) being
a partial mediator of the relationship between CSA and adult depressive
symptoms. The results suggested that a memory of the mother as more caring is
associated with higher levels of depressive symptoms. The recollection of
mother as caring does not moderate the relationship between CSA and adult
depressive symptoms. Variables having to do with the respondent’s memory of
the father-figure (as measured by the PBI) did not reach significance in either
the mediator or moderator analysis. In all significant mediational results,
intervening variables acted as partial mediators, with a direct effect remaining
between CSA and adult depressive symptoms.
2
To Matthew
who is energetically alive in the
hearts and souls of those who love
him.
And to the heart of the appaloosa.
ACKNOWLEDGEMENTS
I would like to thank
Mitch— for helping and respecting me...
Tom— for investing in me...
Heather— for teaching me...
Karen— for being my friend...
Pam— for being my family...
Melissa— for distracting me...
Tasha— for uncomplainingly giving the best years
of her life to this project...
and most of all
The participants in this study—
for sharing their stories.
TABLE OF CONTENTS
Page
LIST OF TABLES AND FIGURES vii
INTRODUCTION 1
Childhood Sexual Abuse . . . . . . 2
Prevalence Among Women: General . 2
Prevalence: Lesbian Women . 4
Long-term Correlates . 5
Factors Influencing the Manifestation and Severity
of Long-term Correlates and Symptoms 7
Models and Mechanisms . 8
Recent Research Focusing on the Relationship
Between Family Environment and CSA Long
-term Correlates . 1 0
Adult Depressive Symptoms . 1 7
Depressive Symptoms in Women in General . 1 7
Depressive Symptoms Among Lesbian Women . 1 7
Depressive Symptoms and Childhood Sexual Abuse 18
Depressive Symptoms and Parental Bonding 25
Summary of Problems in Previous Research 26
Problems in Participant Recruitment . 2 8
Research Design and Questions . 3 1
Proposed Model . 3 2
Research Questions . 3 3
iv
M E T H O D S ..........................................................................................................35
Survey Distribution . 3 5
Measures . 3 7
CES-D and Rosenberg-SEI . . 4 0
Parental Bonding Instrument
Variables Used in This Study . 4 3
Significance Levels . 5 0
Analyses . 5 0
R E S U L T S .......................................................................................................... 54
Survey Returns . 5 4
Descriptive Data . 5 5
Demographic Data . 5 5
Definition and Frequency of Childhood Sexual Abuse 59
Comparison of CSA Group with Non-CSA Group 70
Survey Results . 8 2
Data Analysis/Research Questions . . 9 4
Comparison of Depressive Symptoms . 9 4
Analysis of Variables as Mediators of the Relationship
Between c s a and a d u l t d e p r e s s iv e s y m p t o m s 95
Analysis of Variables as Moderators of the Relationship
Between c s a and a d u l t d e p r e s s iv e s y m p t o m s . 102
Summary: Research Questions .106
v
DISCUSSION
REFERENCES
APPENDIX
LIST OF TABLES AND FIGURES
TABLES
1. Variables Used in the Study . 4 4
2. Comparison of study demographic data with data
from the National Lesbian Healthcare Study and
with U.S. Census Data . 5 6
3. Numbers of participants who reported having
experienced CSA and numbers of participants
who were judged by study criteria to have been
sexually abused as children . . 6 0
4. Mean differences in a d u l t d e p r e s s iv e s y m p t o m s for
respondents who agree and who disagree with the study
judgement about whether they were sexually abused
as children . . . . . . . . 63
5. Mean differences in s e l f -e s t e e m for respondents
who agree and who disagree with the study judgement
about whether they were sexually abused as children 64
6 . Mean differences in m o t h e r -c a r in g for respondents
who agree and who disagree with the study judgement
about whether they were sexually abused as children 65
7. Mean differences in m o t h e r o v e r -p r o t e c t iv e for
respondents who agree and who disagree with the study
judgement about whether they were sexually abused
as children . . . . . . . . 66
vii
Mean differences in f a t h e r -c a r in g for respondents
who agree and who disagree with the study judgement
about whether they were sexually abused as children
Mean differences in f a t h e r o v e r -p r o t e c t iv e for
respondents who agree and who disagree with the
study judgement about whether they were sexually
abused as children . . . . .
Comparison of respondents who reported sexual abuse
with respondents who did not report sexual abuse
(Using A N O V A ) ..................................................
Correlation matrix of variables (excluding
demographics) for all participants
Correlation matrix of variables (excluding
demographic variables and variables having to do
with the nature of abuse) for participants
determined not to have experienced CSA .
Correlation matrix of all variables (excluding
demographic variables) for participants
determined to have experienced CSA according
to the study definition ( c s a = yes) .
Descriptive statistics for survey instruments
(Rosenberg-SEI, CES-D and PBI) .
Comparison of means of the PBI subscales for
groups of respondents reporting CSA histories
with various perpetrators . . . .
Percentages of respondents with CSA histories who
reported sexual experiences with various persons .
Indirect effects for c s a and partial mediators
( s e l f -e s t e e m and m o t h e r -c a r in g ) in the relationship
between c s a and a d u l t d e p r e s s iv e s y m p t o m s
18. Moderator analysis using the s e l f -e s t e e m
variable
19. Moderator analysis using the PBI
variables . . . .
20. Moderator analysis using the variable:
DID YOU TALK TO A THERAPIST?
FIGURES
1 . Schematic model for data analysis .
2. Model proposed for path analysis
3. Results from the path analysis examining
variables as potential mediators in the
relationship between c s a and a d u l t d e p r e s s iv e
SYMPTOMS . . . . .
4. Path analysis in the absence o f the s e l f -e s t e e m
variable . . . . .
INTRODUCTION
The large literature on Childhood Sexual Abuse (CSA) indicates that a great
many women have experienced sexual abuse as children (e.g. Finkelhor,
Araji, Baron, Browne, Peters, and Wyatt, 1986). There also exists a large
literature exploring the high incidence of depression among women in the
general population (reviewed in McGrath, Keita, Strickland, and Russo,
1990). There is a far less voluminous literature on the experiences of lesbian
women. Available information suggests that a larger number of lesbian
women than heterosexual women have experienced CSA (Herman, 1981;
Meiselman, 1978) and that a larger number o f lesbian women than
heterosexual women have experienced significant depressive symptoms as
adults (Rothblum, 1990, 1994).
This introductory section will demonstrate, via literature review, that
CSA has been seen to be a serious problem occurring frequently and having
associated with it serious long-term symptoms. Initially, most research
focused on the prevalence of CSA and description of the numerous long-term
1
correlates of CSA. Recently the focus of research has begun to shift to more
developmental and etiological approaches in an effort to determine what
characteristics of the abuse and/or the environment in which it occurs cause,
mediate, or moderate the long-term correlates. This more recent research was
designed with the intention of trying to explain how and why abuse over
numerous cases appears to be associated with a large number of different
constellations of adult symptoms and characteristics. One of the foci of this
recent research is family environment, and one of the questions within that
focus is whether and to what extent the family environment causes, mediates,
or moderates the ensuing adult symptomatology. The intent of this research
is to explore those relationships with focus on adult symptoms of depression
and on self-esteem. Further, a previously unstudied population— lesbian
women— will be examined. The last part of this introductory section will
frame and summarize the methodology and specific research questions posed
within the study.
Childhood Sexual Abuse
Prevalence Among Women: General
The rate of childhood sexual abuse among women in the general population
in the United States is estimated, by those summarizing the literature, to be in
the vicinity of 20% to 40% (Donaldson and Cordes-Green, 1994) or 15% to
2
45% (Peters, 1988), with individual studies demonstrating rates varying from
a high of 67% (Wyatt, 1985) to a low of 6.8% (Stein, Golding, Siegel, and
Burnam, 1988).
In brief, when looking at a community sample, many different rates of
CSA among women have been reported, including 67% (Wyatt, 1985), 38%
(Russell, 1986), 27% (Finkelhor, 1990), and 23% (Janus and Janus, 1993).
When looking at a college sample, prevalence appears to be at the lower end
of the various rates (e.g. 23% found by Harter, Alexander and Neimeyer,
1988). In psychiatric samples, the rates tend to be generally higher than
those found in community samples. For example, rates reported include 45%
among 78 women outpatients with severe mental illness (Muenzenmaier,
Meyer, Struening, and Ferber, 1993), 44% among 119 women with diagnosed
depressive disorder and 49% among 66 women diagnosed with anxiety
disorder (Murrey, Bolen, Miller and Simensted, 1993). Prevalence findings
also suggest that approximately 50% of sexual abuse is perpetrated by a
parent or parent-figure (Finkelhor and Dziuba-Leatherman, 1994).
Studies from other western countries demonstrate comparable rates of
CSA. In these studies findings include rates from New Zealand of 33%
overall, with 12.5% intrafamilial abuse (Anderson, Martin, Mullen, Romans
and Herbison, 1993) and 13% intrafamilial sexual abuse (Bushnell, Wells,
and Oakley-Browne, 1992); a rate from Canada of 32% CSA in a community
3
sample (Bagley, 1991); and a rate from Switzerland of 12% CSA in a
community sample (Ernst, Angst and Foldenyi, 1993).
The enormous range in prevalence rates appears to be, at least in part,
due to differences in research methodology. The diverse findings seem to be
related to differences in definitions of sexual abuse (e.g., contact v. non-
contact), the way in which questions were asked (e.g., phone, in person,
mail), the types of questions asked (e.g., "were you abused?" v. "did anyone
ever touch you?"), and sampling techniques (e.g., mental health system v.
general population) (Donaldson and Cordes-Green, 1994; Finkelhor et al.,
1986). There appear to be no differences in prevalence of CSA in women
among different racial or ethnic groups, or different socio-economic groups
(Finkelhor et al., 1986; Wyatt, Newcomb and Reiderle, 1993).
Prevalence: Lesbian Women
Suggestions have been made that lesbian women as a group may have
experienced more CSA than women in general (Herman, 1981; Meiselman,
1978). Recently, the National Lesbian Health Care Survey provided data
demonstrating that in their sample of 1,917 lesbian women, 21% had
experienced "rape or sexual attack" as children, and of the 1,779 who
responded to the question about having sex with relatives as children, 19%
said that they had this experience (Bradford, Ryan and Rothblum, 1994).
4
This falls within the established range for CSA and incest in the population
of women in general. However, these numbers may minimize the actual
occurrence of CSA because of the wording of the questions (e.g., women
who were genitally fondled by an uncle may not evaluate the experience as
"rape" or "sexual attack").
Long-term Correlates
There is also general agreement among CSA researchers that childhood
victimization is associated with serious short- and long-term deleterious
effects on children’s mental health (Finkelhor and Dziuba-Leatherman, 1994),
specifically including feelings of sadness and loss of self-esteem (Cerezo and
Frias, 1994). These symptoms have been shown by a number or researchers
to persist into adulthood and to help shape adult mental health and
psychological functioning (e.g., Briere, 1992a; Browne and Finkelhor, 1986;
Courtois, 1988; Herman, 1981, 1992; Roesler and McKenzie, 1994; Tsi and
Wagner, 1978). A wide spectrum of specific long-term correlates has been
identified by numerous researchers over the past decade. These include
PTSD (e.g., Gelinas, 1983); Dissociative Identity Disorder (Multiple
Personality Disorder) (e.g., Kluft, 1985; Putnam, Guroff, Silberman, Barban
and Post, 1986); Borderline Personality Disorder (e.g., Herman, Perry, and
Van der Kolk, 1989; Paris, Zweig-Frank and Guzden, 1993); revictimization
5
(e.g., Herman, 1981; Russell, 1986; Van der Kolk, 1989); substance abuse
(e.g., Boyd, 1993); self-harm and self-cutting (e.g. Van der Kolk, Perry and
Herman, 1991); low self-esteem (e.g., Herman, 1981; Tsi and Wagner,
1978); sexual difficulties (Briere, 1992; Jehu, 1988, 1989); problems in self
development and identity formation (e.g., Cole and Putnam, 1992; Herman,
1981); and problems in social development and interpersonal relating (e.g.,
Cole and Putnam, 1992; Herman, 1981). More recently, a number of other
long-term correlates of CSA have been added to this list, such as increased
susceptibility to HIV infection because of sexual acting out behavior (Allers,
Benjack, White and Rousey, 1993; Bartholow, Doll, Joy, Douglas, Harrison,
Moss and McKirnan, 1994; Cunningham, Stiffman, Dore and Earls, 1994).
This more recent list also includes implication of a CSA history in a number
of physical difficulties, such as chronic pelvic pain (Rosenthal, 1993; Walker,
Wayne, Hansom, Harrop-Griffiths, Holm, Jones, Hickok, and Jamelka, 1992);
pseudo-seizures (Bowman, 1993); gastrointestinal disorders (Drossman,
Leserman, Nachman, Zhiming, Gluck, Toomey and Mitchell, 1990); irritable
bowel symptoms (Longstreth and Wolde-Tsadik, 1993); and possible
disorders of the hypothalamic-pituitary-adrenal axis resulting in a decreased
sensitivity of the hypothalamus to certain chemical messengers (DeBellis,
Chrousos, Dorn, Burke, Helmers, Kling, Trickett and Putnam, 1994; Henry,
1993). — ~
6
In addition to these correlates, there are numerous researchers who
have associated adult affective disturbances-especially depression--with a
history of childhood sexual abuse (e.g., Allers, Benjack and Allers, 1992;
Bifulco, Brown and Adler, 1991; Briere, 1992a, 1992b; Browne and
Finkelhor, 1986; Garnefski, van Egmond and Straatman, 1990; Gelinas,
1983; Herman, 1992; Koverola, Pound, Heger and Lytle, 1993; Tsi and
Wagner, 1978). In turn, loss of self-esteem has been implicated in the
development of depression (Jack, 1991; Jackson, 1986). Researchers in the
area of trauma have linked depression to feelings of self-hatred and decreased
self-worth (McCann & Pearlman, 1990; Terr, 1990). The precise relationship
between the CSA, self-esteem or self-worth and depression, however, has not
been elucidated.
Factors Influencing the Manifestation and Severity
of Long-term Correlates and Symptoms
The CSA literature discusses aspects of the experience of sexual abuse that
influence the type and severity of the long-term effects potentially
experienced by the person who was abused. These include number of
perpetrators, relationship to the perpetrator, duration and/or frequency of
abuse, violence or force used, type of sexual act, age of perpetrator
7
(Finkelhor et al., 1986; Wyatt, Newcomb and Riederle, 1993). In addition, at
least one factor has emerged that might ameliorate the effects of the abuse:
whether the child was able to tell someone and be supported and believed.
Aspects of the abuse that have been seen to be associated with more
intense or worse long-term correlates include the use of force, prolonged
abuse, penetration, and abuse by a primary caretaker (Finkelhor et al., 1986;
Herman, Russell and Trocki, 1986; Russell, 1986; Wyatt, Newcomb and
Riederle, 1993). These abuse characteristics were also seen to be important
correlates of abuse outcome in a review of 45 studies by Kendall-Tackett,
Williams and Finkelhor (1993). In addition, these reviewers cited maternal
support as an important correlate of outcome. Among those studying the
relationship between CSA and adult depression, Roesler and McKenzie
(1994) found that abuse by the father and use of physical force or verbal
intimidation were associated with adult depression. In addition, they found
the use of force to be correlated with decreased self-esteem in adulthood.
Models and Mechanisms
A number of researchers have noted that there does not seem to be a simple
relationship between CSA and adult psychological symptoms (e.g.,
Alexander, 1992; Yama, Tovey and Fogas, 1993). The number of correlates
8
associated with CSA is, as previously shown, very large— and adults who
have experienced CSA may have a different number and combination of
symptoms. Initially research in the field was descriptive, and generally
limited to investigating the prevalence, nature and effects of abuse. Some
researchers, however, have focused— generally theoretically-on how the
experience of abuse might be transformed into the observed long-term
correlates.
A number of developmental models have been proposed. These
include the "child abuse accommodation syndrome" described by Summit
(1983), the "traumagenic" model proposed by Finkelhor and Browne (1985),
the "information processing of trauma" model of Hartman and Burgess
(1994), and "constructivist self-development theory" o f McCann and
Pearlman (1990). These and a number of other less formal mechanisms
proposed to explain the relationship between CSA and its long-term correlates
are described elsewhere (Donaldson and Cordes-Green, 1994; Neumann,
1994).
Formation of long-term symptoms has been proposed to include
mechanisms such as cognitive distortions (Celano, 1992; Gold, 1986; Harter,
Alexander and Neimeyer, 1988); learning history (Jehu, 1988); simultaneous
emotional abuse integral to the sexual abuse (Briere, 1992a); invocation of
psychological defenses (Blake-White and Klein, 1985; Courtois, 1988);
9
truncated development (Alexander, 1992; Meiselman, 1978); and factors in
the family environment (Alexander, 1985 & 1992; Bushnell, Wells and
Oakley-Browne, 1992; Conte and Schuerman, 1987; Edwards and Alexander,
1992; Peters, 1988). This is only a representative sample of researchers and
the suggestions that have been made to explain the transformation of event to
outcome.
Recently Alexander (1992) has suggested that attachment (or parental
bonding) may have a very important effect on the long-term correlates of
CSA, mediating (or moderating?) the relationship between these variables.
Alexander states:
...failure to consider the relationship context of sexual abuse ignores
an important aspect of the long term effects. (1992, p. 185)
There have been, however, no studies to date looking at
attachment specifically in a sexually abused population. (1992,
p. 186)
There are also suggestions from the literature of Self Psychology that such
attachment may provide the developing person with selfobjects that moderate
the effects of CSA (Neumann, 1991; White and Weiner, 1986). In spite of
these suggestions, the research literature to date focuses on the family in
general and family structure rather than relationships with and attachment to
parental figures.
10
Recent Research Focusing on the Relationship Between
Family Environment and CSA Long-term Correlates
In the past two years some research studies have been published reporting
investigation of family influences. Nash, Hulsey, Sexton, Harralson, and
Lambert (1993) suggested that families in which abuse occurs seem to be
more pathological than non-abusing families, with higher levels of role
confusion, boundary confusion and rigid behavioral control combined with
less cohesiveness and adaptability. Nash et al. (1993) make an interpretation
which sounds very much like a suggestion that it is actually the family
situation that is responsible for long-term correlates of CSA rather than and
exclusive o f the abuse itself:
... recent empirical work suggests that some adult pathology
associated with childhood sexual abuse may reflect the effects
of a broadly pathogenic home environment rather than those of
the abuse per se. (Nash et al., 1993, p. 277).
A similar position is taken by Pribor and Dinwiddie (1992) who observe that
incest might result in trauma which leads to adult psychological difficulties,
or that incest could be a "marker" for a severely dysfunctional family:
It may be that the dysfunctional family environment rather than
the incestuous act per se plays a major role in the genesis of
interpersonal problems leading to the development of
psychiatric symptoms and disorders. (Pribor and Dinwiddie,
1992, p. 55).
A somewhat different approach is taken by other researchers who view the
family environment as influential in determining risk and outcome of CSA,
but not as exclusively causative. Along these lines, is research by Wyatt and
Newcomb (1990) who reported on a study of 111 women from a community
sample using retrospective reports of CSA. They examined 11 variables
grouped into three categories (circumstances of abuse; mediators of abuse;
and negative effects of abuse), and found direct effects by two circumstances-
of-abuse variables (severity of abuse and proximity o f abuse). Other
circumstances-of-abuse variables predicted mediator variables which, in turn,
appeared to influence negative outcome. These mediator variables were
"Immediate negative responses," "Internal attributions," "Extent of
disclosure," and "Involvement of authorities." In their comments about the
research, Wyatt and Newcomb caution us that it is necessary to look for
mediator processes and moderators, and to distinguish (in the manner of
Baron and Kenny, 1986) between the two.
Long and Jackson (1994) studied 172 female college students, 80 of
whom stated that they had experienced CSA. In this study families were
examined using reports of the participants on the Family Environment Scale
and classifying the families according to the system of Billings and Moos
(1982). This system divides families into functional and dysfunctional types.
Subgroups under functional families are: Independence-oriented,
intellectual/cultural-oriented, and support-oriented. Dysfunctional family
12
types are: conflict-oriented, disorganized; achievement-oriented; structured
moral/religious-oriented; and unstructured moral/religious-oriented.
Essentially the same number of women with and without CSA
histories were found by Long and Jackson (1994) in all family types except
two: the support-oriented families had significantly fewer women in the CSA-
history group, and the disorganized-family group had significantly more
women in the CSA-history group. The conclusion in the study was that there
are perceived differences in family functioning which can, in part, determine
differences in risk for CSA. This may be the case, with three caveats. First,
the data are retrospective and depend on the participant’s reported memory of
her family. Second, the number of women in each of these groups was
small, with, for example, 16 women in the support-oriented family group and
6 women in the disorganized family group. Third, the opposite interpretation
is also possible. That is, although the risk was seen to be higher in one
dysfunctional family type and lower in one functional family type, in most
family types (both functional and dysfunctional) the risk was essentially the
same with roughly equal numbers of women having experienced or not
experienced CSA.
Similarly, Yama, Tovey and Fogas (1993) proposed that family
environment might be important in either of two ways: by having direct
effects independent of the CSA, or by interaction with the CSA to produce
13
the long-term effects. The study looked at 46 women with histories of CSA
and 93 with no such history. The Moos Family Environmental Scale was
also used in this study to classify families of origin, together with measures
of adult depression and anxiety. The authors concluded that, in their sample,
anxiety and depression were directly related to the level of conflict perceived
in the family of origin, and were inversely related to the perception of
cohesion and independence. CSA appeared to be related to higher levels of
depression in families perceived to be high in conflict and in families
perceived as cohesive. Alternatively, CSA appeared to be associated with
lower levels of depression when families were perceived as high in control.
Scores on anxiety and depression measures were not reported in this study, so
it is unclear what level of symptoms were involved (i.e., were the differences
between being very depressed and moderately depressed, or between having a
few and no symptoms at a level below the cut-off for depression?). Still, this
is another indication that family functioning may have some kind of effect
and predictive power.
The view that family environment is central to CSA risk and outcome
was extended by Harter, Alexander and Neimeyer (1988) who did a study
with 29 college women who had experienced CSA and 56 college women
who had not experienced sexual abuse (although they may have experienced
physical abuse or neglect). These researchers measured social adjustment as
14
their outcome, or dependent variable, using the Social Adjustment Scale.
This measure consists of an interview assessing functioning in employment,
school or "housewife" roles; social and leisure activities; extended family
relationships; marital and significant other relationships; and parenting. The
sexually abused participants in this study reported less cohesive and adaptable
families of origin. These researchers found that family characteristics and
increased perception of social isolation were more predictive of social
maladjustment in adulthood than the sexual abuse per se. There was one
exception to the finding that the sexual abuse was more predictive, and that
was in cases of sexual abuse by a father-figure. In that instance, the CSA
was related to poor social adjustment even after family and social-cognitive
variables were controlled.
A slightly different view of sexual abuse trauma is offered by Green
(1988) who sees it as superimposed on chronic pathological family
functioning, with each contributing different elements to adult
symptomatology. He proposes that the CSA is associated with feelings of
powerlessness, fearfulness and anxiety and with anxiety-related symptoms
such as nightmares, sleep disturbances, hypervigilance and dissociation.
These symptoms are characteristic of PTSD, and the logical extension would
be that the abuse may be associated with formation of that syndrome. Green
proposes that family characteristics are associated with stigmatization,
15
betrayal, guilt, shame, low self-esteem, mistrust, depression and protracted
use of pathological defenses. The extension of this, similarly, would be that
the long-term affective sequelae associated with CSA issue from the family
characteristics in which the CSA is imbedded.
Finally, in a study of 17 women who had experienced parental incest
Wind and Silvern (1994) found that parental warmth was a strong influence,
affecting the relationship between intrafamilial sexual abuse and adult
depression and levels of self-esteem. The major difficulties in this study
appear to be the small number of subjects as well as some uncertainty on the
part of this reader about the exact meaning of terms such as "mediate" in the
paper. For example, the first two paragraphs of the introduction discuss the
relationships between CSA, unsupportive parenting, high family stress and
adult correlates of CSA using all of the following terms: exacerbate, account
for, attenuate, influence, mediate, affect and ameliorate. It is not clear when
and if distinctions are being made between cause, mediation and moderation.
In summary, suggestions have been made that attachment (or parental
bonding) may be an important mediator affecting the long-term correlates of
CSA. Research to date has looked at a number o f family-related variables as
potential mediators or moderators of CSA. Included are studies looking at
family type or broad characteristics of family structure, with only one study
looking at "parental warmth." To test the theoretical formulations around
16
attachment and bonding, a study must be devised using an instrument or
method designed to tap into the relationships between parents and the child.
Adult Depressive Symptoms
Depressive Symptoms in Women in General
The recent publication of the APA Task Force on Women indicates that
depression is a significant problem for women in general, with the rate of
depression in women being twice the rate in men (McGrath et al., 1990). A
great many factors have been suggested or shown to be associated with
depressive symptoms. These include feelings of helplessness or hopelessness
(Abramson, Garber, and Seligman, 1980), early sorrow or loss (Gurian, 1987)
loss of self-esteem (Jack, 1991; Jackson, 1986), and internalized anger (Jack,
1991; Lemer, 1987). The APA task force report also lists a number of
factors thought to "elevate risk" for depressive symptoms in women. These
focus largely on the more immediate, proximal factors associated with
depressive symptoms, such as women’s roles and status, poverty, and
diagnostic bias.
Depressive Symptoms Among Lesbian Women
As a result of clinical observation, it was suggested by Rothblum (1990) that
the rate of depression among lesbian women may be higher than that of
17
women in general. However, studies of lesbian women and depression are so
scarce that, as Rothblum (1994) points out, of the several hundred studies
reviewed in the Task Force Report of the American Psychological
Association on women and depression (McGrath et al., 1990), not one paper
focused on the experiences o f lesbian women. Recently, with the first
publication of data from the National Lesbian Health Care Study, come the
first estimates of depression in the lesbian population, with over 33% of the
1,917 lesbian women surveyed stating that they had experienced a "long
depression or sadness" at some time in the past, and 11% stating that they
were experiencing depression currently (Bradford, Ryan and Rothblum,
1994). Other research indicates that, compared to heterosexual women,
lesbian women have higher rates of suicide and alcoholism— both of which
have been shown to be associated with depression (Erwin, 1993; Rothblum,
1990, 1994). Depression and depressive symptoms, then, are of considerable
concern in the lesbian community, but in spite of that, they have not yet been
studied.
Depressive Symptoms and Childhood Sexual Abuse
The depression literature has recently been criticized for a paucity of
studies examining the relationship between childhood trauma (and specifically
CSA) and later development of depressive symptoms or depression (Cutler
18
and Nolen-Hoeksema, 1991). Such traumatic events and their accompanying
feelings would be expected to be influential and important in the development
of personality and adult affect (Herman and Van der Kolk, 1987; Terr, 1990).
Since Cutler and Nolen-Hoeksema’s (1991) criticism of the literature,
a few studies have included a quantitative look at adult depression or
depressive symptoms as a correlate of CSA. For example, Fox and Gilbert
(1994) have recently studied the long-term effects of childhood physical
abuse, incest and parental alcoholism, using the Beck Depression Inventory
(BDI) as one of their measures of adult psychological functioning.
Participants in their study were 253 college women with a mean age of
19.33. In this group they found that women who had experienced more than
one of the three childhood traumas (physical abuse, incest and parental
alcoholism) scored significantly higher on the BDI than those who had
experienced only one or none of these traumas. The authors also observed
that few respondents experienced only one of these traumas, with most
having experienced none or more than one. Scores on the BDI were said to
have been in the "mild" range for people who experienced one type of
trauma, and in the "moderate" range for those who experienced more than
one trauma.
A similar finding is reported by Wind and Silvern (1992) who present
data supporting the view that depression is a long-term correlate of CSA, and
19
demonstrating an association between both physical and sexual abuse and
depression. Their conclusions include the proposition that various kinds of
abuse and trauma may be linked to adult depression and that different types
of abuse often co-occur.
Levels of adult depression in women who had or had not experienced
a recent trauma were examined in another study by Gidycz, Coble, Latham
and Layman (1993). These women were divided into four groups: women
who had experienced a trauma recently and who either did or did not have a
history of CSA (groups 1 and 2), and women who had not experienced a
trauma recently and who either did or did not have a history of CSA (groups
3 and 4). In this study respondents were 857 college women. Criteria used
to define a history of CSA were not specified in the paper. Analysis using
ANOVA shoed that these groups had significantly different scores on the
BDI with post hoc analysis demonstrating that (1) those who had recent
trauma experiences and a history of CSA scored significantly higher on the
BDI than those with no recent trauma, and (2) among respondents who
reported no recent trauma, those who had a history of CSA scored
significantly higher on the BDI than those with no history of CSA. In other
words, recent victimization in combination with a history of CSA was
associated with more depressive symptoms in the adult women when
compared to women with no recent victimization experiences. In addition,
20
among those who had not been recently victimized, a history of CSA was
associated with more adult depressive symptoms.
In an attempt to examine the spectrum of etiological agents potentially
implicated in major depression in adult women, Kendler, Kessler, Neale,
Heath, and Eaves (1993) surveyed 680 female twin pairs. The dependent
variable in their study was the DSM-III-R diagnosis of a major depressive
episode within the past year. Predictors were: (1) genetic factors, (2) parental
warmth, (3) childhood parental loss, (4) lifetime traumas, (5) neuroticism, (6)
social support, (7) past depressive episodes, (8) recent difficulties, and (9)
recent stressful life events. A path analytic approach was used, and all nine
variables together accounted for a reported 50.1% of the variance in the
dependent variable. The authors reported that the largest effects were seen
for (in descending order) recent stressful life events, genetic factors, past
depressive episodes, and neuroticism. The concluded that traumatic
experiences, genetic factors, temperament, and interpersonal relationships
comprise the "interacting risk factor domains needed to understand the
etiology of major depression" (p. 1139).
These authors pointed out a number of important limitations to their
study. A number of additional serious questions can be raised. First, the
defining characteristics of the independent variables must be examined. For
example, depression was said to be characterized by genetic factors if the
21
subject’s twin had also experienced at least one episode of major depression
at some time in her life. But, because one’s twin has experienced depression
does not necessarily make the depression genetic, nor would it identify
"genetic" depressions in which the subject’s twin had just not had her
depressive episode yet (mean age was 30.3). Similar environments could be
involved also. Participants were all raised in the same household until at
least age 16. Second, these authors used the Parental Bonding Instrument
(PBI) to measure parental warmth, however, they averaged the scores for
twin pairs. They also averaged each twin’s response for each parent yielding
an average score for both parents as evaluated by both twins. There appears
to be no precedent in the literature for this method, and there is literature
suggesting that the PBI scores relating to the mother have a significant
association with adult depression, while the scores for the father do not
(Bifulco, Brown and Harris, 1994; Parker, 1983). Third, a construct like
"neuroticism" may, perhaps, be a result of the same factors that contribute to
the development of depression, rather than antecedent to the depression. In
short, the paper has limitations serious enough that the results cannot be
confidently interpreted.
The path analytic approach taken by Kendler and associates (1993) to
examine the question of etiologic agents for depression is an important
contribution. This type of analysis allows a more comprehensive view of a
22
variety of variables and their relationships to depression. Of similar
importance is the path analytic approach to the question of mediators and
moderators of sexual abuse explicated by Wyatt and Newcomb (1990) and to
the question of predictors of adult sexual behavior explicated by Wyatt,
Newcomb and Riederle (1993). These studies suggest that such an approach
might be useful in the exploration of the mediators and moderators involved
in the relationship between childhood sexual abuse and adult depression.
A different approach to questions about CSA and depression was
taken by Valentine and Feinauer (1993) who studied 22 women who had
experienced sexual abuse in childhood and who, at the time of the study,
were described as having "relationships, stable careers, and healthy
personalities" (p.216). Interviews were used in an effort to determine what
aspects of their lives the participants thought had given them resilience or
helped them to form a positive outcome following a negative (CSA)
experience. The themes extracted from these interviews included the ability
to find positive relationships, positive self-regard, spirituality, external
attribution, a sense of personal power, and a positive view of life. One
hastens to point out that these themes could be seen as characteristics of the
woman who has coped successfully, and are not necessarily explanatory of
the structure within which such successful coping developed. For example,
there is no way of knowing from this design whether the woman’s external
23
attribution was an in-place characteristic enabling her to cope with the CSA
experience, or whether, in the process of coping and making sense of her life,
she developed a more external attrbutional style.
In summary, these studies add support to the connection make in
earlier child abuse literature between CSA and adult depression. In addition,
they provide data suggesting that more than one trauma or type of abuse may
frequently be associated with adult depression, and that CSA may be just one
of a number of childhood traumas (including physical abuse) associated with
adult depression. Most of the studies examined here use methods involving
comparison of means or looking at the relationship between variables and
adult depression one variable at a time. However, in one case (Kendler et al,
1993) a path analytic approach was employed which allows for consideration
of the roles of a number of variables at the same time. This type of approach
has been taken by researchers studying related constructs (Wyatt and
Newcomb, 1990; Wyatt, Newcomb and Riederle, 1993). Such an approach
would seem to accommodate a greater number of potentially important
variables and more closely approximate the complexity of the relationships
between numerous variables.
24
Depressive Symptoms and Parental Bonding
A small number of papers have looked at the relationship between adult
depression and parental characteristics such as warmth, caring and control.
The earliest of these employed a retrospective, self-report measure named the
Parental Bonding Instrument (PBI). In their review of early studies with the
PBI, Burbash, Kashani and Rosenberg (1989) concluded that depressed adults
remember their parents as exhibiting low care and high over-protection as
measured by the PBI. Parker (1983) had previously labeled this combination
"affectionless control" and suggested that it may play an important role in the
development or maintenance of adult depressive disorder. More recent data
do not support the notion that "affectionless control" is specifically associated
with depression in adolescents, but that it is a correlate of adolescent affective
and behavioral problems in general (Burbash et al., 1989) and adult
Borderline Personality Disorder (Zweig-Frank and Paris, 1991).
Along these lines, the PBI and a newer instrument called the
Childhood Experience of Care and Abuse interview (CECA) were both used
by Bifulco, Brown and Harris (1994) to assess the relationship between adult
depression and quality of care and relationships with parents. In this study,
other variables including physical and sexual abuse were also tested. In both
samples studies (225 and 395 women) parental indifference and antipathy
from the mother were seen to be most strongly related to adult depression.
25
Parental control was not seen to be significant unless high parental
indifference was first eliminated, and the authors pointed out that there is
some inconsistency in the literature as well as in their data about the role of
parental control or over-protection. More work needs to be done in this area
in general, and, again, no work had been done in the lesbian population.
Summary of Problems in Previous Research
Previous research has failed to address issues of family dynamics and
parental bonding in relation to CSA and its effects (Alexander, 1992). In
addition, research has ignored the experience of lesbian women. Virtually no
studies have been done investigating CSA, parental attachment, or depressive
symptoms in this group. The self-esteem of lesbian women has been
discussed from a clinical point of view, mostly in a speculative way with
regard to either the stigma of belonging to a minority group (e.g. Birt and
Dion, 1987; Christie and Young, 1986; Martin and Hetrick, 1988) or in
relation to formation of a lesbian identity (e.g. Padesky, 1989; Sophie, 1987).
In the existing CSA literature it is sometimes difficult to discern
whether people are talking about cause, mediation, or moderation when
discussing or studying the relationship between CSA variables and long-term
outcome. A variable such as family environment could, potentially, be a
cause by directly producing the observed outcome, independent of abuse.
26
Family environment might, alternatively, mediate outcome, in which
case it would act as a transformer, intervening in the relationship between the
independent and dependent variables For example, a mediating process
might take the initial output (CSA) and transform it into different outcome or
symptoms (perhaps depression v. aggression, hypothetically) depending on
the different qualities of the family environment (perhaps denial v. aggressive
retaliation— again, hypothetically).
Finally, family environment could potentially be a moderator by
changing the direction or strengths of the relationships between the CSA and
the potential long-term correlates. For example (hypothetically), increasing
amounts of denial of CSA in a family might be associated with increasing or
decreasing amounts of depression in the adult children.
These definitions, explicated by Baron and Kenny (1986) will be used
in this paper. It is suggested here that some of the confusion in CSA
research stems from inconsistent or imprecise use of these terms.
A problem also exists with the separation between the CSA and
depression literatures, reducing the possible impact of CSA research findings
on the design and development of research on depression in women. This
has, as stated previously, been recognized by Cutler and Nolen-Hoeksema
(1991) who have suggested that CSA should be included in the variables
27
examined in studies on women and depression. This inclusion has only
recently begun.
Finally, as pointed out by Rothblum (1994), there is virtually no
research on lesbian women in any literature. This indictment is certainly true
in the literatures of depression and childhood abuse of all sorts.
In short, within the CSA literature, depression has been shown to be
one of the long-term correlates of CSA (e.g. Briere, 1992a&b, Herman, 1992)
and the scant literature focusing on lesbian women suggests that high rates of
depression and CSA may be found in this group.
It, therefore, seems appropriate to investigate CSA in the lesbian
population, both to determine whether the prevalence there is indeed higher
than in the general population of women, and because there is some reason to
believe that the lesbian population will provide more respondents who have
experienced CSA.
Problems in Participant Recruitment
One of the difficulties in research focusing on lesbian women, as pointed out
by Rothblum (1994) has to do with selection or recruitment of participants.
According to Rothblum, there are three major methods of obtaining
participants: (1) by membership in gay and lesbian organizations, (2) by self-
28
identity as lesbian or gay, and (3) by involvement in same-gender sexual
activity. Indeed, there is currently no real consensus about what the
definition of lesbian is. Is a lesbian woman one who thinks of herself as
lesbian, or one who behaves as a lesbian by having sexual contact with
another/other woman/women? These questions are addressed more
extensively in the literature on lesbian identity and lesbian identity formation
(e.g., Cass, 1979, 1983-84; Minton and McDonald, 1983-84; Troiden, 1988).
Whatever method of participant selection is used, certainly a number
of people will be left out. This is not news to psychological research in
general, and has been a much-discussed problem in research not limited to
lesbian women (see Rosenthal and Rosnow, 1991). In the case of this study,
the method of obtaining participants was based on Rothblum’s second
possibility: women who self-identify as lesbian, no matter what their sexual
behavior or their organizational affiliations. This was thought to be the
broadest possible criterion, including women who thought of themselves as
lesbian but had not (or had not yet) acted on this sexually and women who,
for whatever reason, were not affiliated with lesbian organizations as well as
those who were.
Solicitation o f participants via lesbian/feminist newspapers increased
the possibility that subscribers in rural locations, women who were isolated
from the lesbian community and women who were economically
29
disadvantaged or incarcerated might have a chance to participate. Other
methods of solicitation used in this study focused on urban areas and
concentrations of lesbian women, such as the gay-pride celebration in
Pasadena, CA or the distribution of survey forms via bookstores. With one
exception, the bookstores were located in urban centers, the exception being
the store in Provincetown, MA which attracts women from the eastern half of
the country as a vacation center (thus limiting the sample there to more
affluent, but possibly including more rural women).
In summary, caveats must be made about the bias in recruitment of
participants in this study: they were recruited in ways that may leave out the
less affluent and the more rural and isolated members of the lesbian
community, as well as those who have not identified themselves as members
o f the lesbian community. It is true that these samples are biased, but since
virtually no research has been done in the area of lesbian psychology it seems
best to go ahead in spite of the problems and at least start somewhere. In the
words of Kaplan:
On the one hand, the sample is of no use if it is not truly
representative of its population, if it is not a "fair" sample. On
the other hand, to know that it is representative, we must know
what the characteristics of the population are, so that we can
judge whether the sample reflects them properly; but in that
case, we have no need of the sample at all. (Kaplan, 1964,
p.239.)
Rothblum is more specific about the use of participants from the gay and
30
lesbian communities who are self-identified and "out," saying that although
such samples are often criticized for bias, there is much to be gained from
studying lesbian- and gay-identified people. They are the most visible both
within and outside of their community, and thus the focus of attention as
prototypes, role-models, victims of hate crimes, and representatives of their
community at large. In addition, they may be in some ways the most
supported and well-adjusted members of the community, but also the most
stigmatized and stressed. They are definitely worth studying (Rothblum,
1994).
Research Design and Questions
The current study was designed to look at self-esteem, adult symptoms of
depression, and parental bonding in a population of lesbian women who
reported having experienced CSA, and a comparison group of lesbian women
who did not report a history including CSA. This was done to determine
whether either self-esteem or the quality of attachment to parental figures
could be shown to be acting as a mediator or a moderator with regard to the
adult experience of depressive symptoms. In addition, use of
psychotherapeutic services was examined to see if it acts as a moderator of
levels of adult depressive symptoms.
31
Proposed Model
The first part of the study describes the demographics of the sample, and then
delineates rates of CSA; levels of adult depressive symptoms, parental
bonding, and self-esteem; and use o f psychotherapeutic services in a sample
of lesbian women.
The second part of the study concerns relationships between these
variables (CSA, self-esteem, parental bonding, adult depressive symptoms,
and use of psychotherapeutic services) in a sample of lesbian women. This
second part o f the study specifically examines CSA as one correlate of
depressive symptoms in adult lesbian women. Self-esteem is proposed in this
study to be a mediator, with CSA being associated with lowered self-esteem,
which, in turn, is associated with adult depressive symptoms.
Parental bonding has also been suggested to mediate the relationship
between CSA and adult depressive symptoms (Alexander, 1992). This study
examines both parental bonding and self-esteem as potential mediators or
moderators of the relationship between CSA and adult depressive symptoms.
Finally, another moderator which might mitigate the effects of CSA in
adulthood (participation in psychotherapy) is investigated. The suggested
relationships between these variables are illustrated in Figure 1.
32
Research Questions
The first part of this study asks the questions: (1) Who are the people in the
study? and (2) What is the experience of lesbian women with regard to
childhood sexual abuse, self-esteem, parental bonding, depressive symptoms
and use of psychotherapeutic services?
The second part of the study addresses four questions:
1) Are lesbian women who report having been sexually abused in
childhood more likely to show depressive symptomatology in
adulthood than lesbian women who did not report that they were
sexually abused as children?
2) Does self-esteem mediate or moderate the relationship between CSA
and adult depressive symptoms?
3) Does the respondent’s retrospective report of parental bonding mediate
or moderate the relationship between CSA and adult depressive
symptoms?
4) Does participation in psychotherapy moderate the relationship between
CSA and adult depressive symptoms?
33
Figure 1 . Schematic model for data analysis, showing the conceptual
possible relationships between variables in this study
POTENTIAL
MEDIATOR
(INDEPENDENT VARIABLE:
SELF-ESTEEM or
PBI SUBSCALES)
INDEPENDENT DEPENDENT
VARIABLE DIRECT EFFECT w VARIABLE
(CSA) (ADULT DEPRESSIVE
SYMPTOMS)
POTENTIAL
MODERATOR
(INDEPENDENT
VARIABLE: DID YOU
TALK TO A THERAPIST?)
34
METHODS
Survey Distribution
A total of 929 survey forms were distributed in three ways. First, a number
of forms (129) were accepted by passers-by at a booth stationed at a "coming
out day" celebration in Pasadena, CA. Fifty-seven respondents filled out the
form at the event and returned it immediately, while 53 returned the forms
through the mail at a later time (totaling 85% returned).
A second group of forms (599) were distributed via lesbian/feminist
bookstores. The first 259 of these were distributed by prior arrangement with
two bookstores, one in Pasadena, CA and the other in Provincetown, MA.
An encouraging number of the initial bookstore surveys were quickly filled
out and returned, and so another 340 were mailed to 17 lesbian/feminist
bookstores throughout the country' without prior arrangement with the store
owners. O f the 599 forms distributed via bookstores, a total of 232, or 39%
were filled out and returned by participants. No forms were returned from
1 Store locations: Anchorage, AK; Eugene, OR; New Orleans, LA; Atlanta, GA;
Minneapolis, MN; Madison, WI; Washington, DC; Northampton, MA; Cambridge,
MA; New York, NY; Philadelphia, PA; Cincinnati, OH; Indianapolis, IN; Houston,
TX; Denver, CO; Seattle, WA; Oakland, CA.
35
the vicinities of Minneapolis or New Orleans, suggesting that these stores
may not have made the forms available to their customers. Finally, ads were
placed in two lesbian/feminist papers (Sojourner, published in Cambridge,
MA and Lesbian News, published in Los Angeles, CA) requesting that people
willing to fill out my survey return the ad with their address, and a form
would be mailed to them (complete text is in the Appendix). These
publications were selected because of their large circulation (65,000
combined), and their ability to reach women in a variety of settings, both
geographically and in terms of the types of places women would be likely to
pick up the paper. For example, both have subscribers, but are more often
obtained at newsstands, feminist bookstores, and women’s centers, with the
Lesbian News distributed free of charge. In addition, both are available free
of charge to women’s correctional institutions, shelters and community
centers. The hope was to get as broad a sample as possible. A total of 201
requests for survey forms were received. Forms were mailed to those 201
individuals, and returned by 151 of them (or 75%). Requests for survey
forms came from women in 24 states, Canada, Puerto Rico, U.S. Military
FPO, and U.S. women’s correctional institutions.
An overall response rate is impossible to calculate. Of 929 forms
distributed, 493 (or 53%) were returned. However, this does not take into
consideration the number of people who did not accept a form at "coming out
36
day," or who may have picked one up at a bookstore and put it back on the
shelf. In addition, some participants wrote that they had xeroxed the form
and given it to friends, so the precise number of forms actually in circulation
is not known.
Measures
Each survey package contained, in this order, a page of demographic
questions, the Rosenberg Self-Esteem Inventory (Rosenberg-SEI), the Center
for Epidemiological Studies-Depression Scale (CES-D), the Parental Bonding
Instrument (PBI), six questions about parental figures, an Early Sexual
Experiences Questionnaire, and several open-ended questions about
depression and about sexual orientation. This survey package can be found
in the Appendix. The PBI, CES-D and Rosenberg-SEI are standard, self-
report measures of levels of parental bonding, depressive symptoms and self
esteem respectively. Specific information about these instruments is given in
the paragraphs which follow.
CES-D and Rosenberg-SEI
The CES-D is a twenty-question scale answered on a 4-point Likert
scale (0 to 3) with higher scores indicating more depressive symptoms. The
CES-D has been evaluated extensively for reliability and validity, and has
37
been shown to be internally consistent (Cronbach’s alpha=.84 to .90) and
valid by a number of means (Devins and Orme, 1985). It is composed of
questions which have been variously divided into four areas: negative affect,
positive affect, somatic symptoms, and decreased activity (Hsu and Marshall,
1987); and depressed mood, psychomotor retardation, lack of well-being, and
interpersonal difficulties (Gatz and Karel, 1990). Scores range from 0 to 60,
with a cutoff established at 16 for depressive symptomatology. Scores above
16 indicate presence of depressive symptoms (Devins and Orme, 1985 ; Hsu
and Marshall, 1987). Some researchers have additionally used the following
ranges of scores to indicate different levels of depressive symptoms: 0-15,
no symptoms; 16-20, mild depressive symptoms; 21-30, moderate depressive
symptoms; 31 and higher, severe depressive symptoms (Hsu and Marshall,
1987).
The Rosenberg-SEI is composed of 10 questions, answered on a 4-
point Likert scale (1 to 4). The range of possible scores is from 10 to 40,
with higher scores indicating lower self-esteem.2 In the work presenting his
self-esteem instrument, Rosenberg differentiates between self-esteem and
depression as constructs and shows that his self-esteem inventory is a
unidimentional (single-factor) scale (Rosenberg, 1965). Subsequently,
2 The Rosenberg-SEI is usually scored from 10 to 40 with higher numbers
representing higher self-esteem. It was reverse-scored for the purposes of this
research.
38
investigators have analysed the instrument with some conculding that the
Rosenberg-SEI is, indeed, unidimensional (e.g., O’Brien, 1985) and some that
it is bi-dimentional (Kaplan and Pokorny, 1969; Shahani, Dipboye and
Phillips, 1990). The researchers finding the Rosenberg-SEI to be bi-
dimensional stated that it was measuring both "self-enhancement" and "self-
derrogation," and suggest that the latter construct may be related to
depression.
The idea that the CES-D and Rosenberg-SEI were measuring the same
construct was tested in a study using factor analysis and regression to
examine the relationships between the CES-D and measurement of anxiety
(using Spielberger’s State-Trait Anxiety Inventory, or SSTAI) and the
measurement of self-esteem (using the Rosenberg-SEI) (Orme, Reis, and
Herz, 1986). Their findings showed that CES-D items did not load with
Rosenberg-SEI or SSTAI items in factor analysis. In addition regression
showed a much larger overlap between trait anxiety (which accounted for
49% of the variance in the CES-D scores) than self-esteem (which accounted
for an additional 8% of the variance in CES-D scores, or state anxiety (0.6%
of the variance).
Because of conflicting conclusions in the literature, and in an effort to
proceed cautiously in this research, a second CES-D total ( c e s -d a b b r e v ia t e d )
was computed using only 17 of the 20 items. The three CES-D items
39
eliminated were judged, based on face validity, to be tapping the same
information as questions on the Rosenberg-SEI, thus increasing, by definition,
the association between scores on the two measures. The eliminated CES-D
items were: "I felt that I was just as good as other people," "I thought my
life had been a failure," and "I felt that people disliked me."3 The Pearson
correlation between the Rosenberg-SEI total and the total CES-D score was
.730 (p < .001). Between the Rosenberg-SEI and the shortened CES-D the
correlation was .694 (p < .001).
Parental Bonding Instrument (PBI)
The PBI is comprised of two 25-question scales (one for mother and
one for father) with responses given on a 4-point Likert scale (0 to 3).
Responses are scored yielding four subscales consisting o f a "caring" subscale
and an "over-protection" subscale for each parental figure. PBI questions and
scoring are shown in the Appendix.
The PBI was developed in an effort to "examine the parental
contribution to a parent-child bond and to attempt to define and measure the
3 A factor analysis with these data and using the Rosenberg-SEI and CES-D items
together (4 factors, varimax rotation) demonstrated that the first two of these
eliminated CES-D questions loaded on a factor with the Rosenberg-SEI items and
were the only CES-D items to load on that factor. The third eliminated CES-D
question loaded on a factor with other CES-D items but no Rosenberg-SEI items.
The removal of the third question may, therefore, have been unnecessary, but was
shown to have little effect on the value of statistics, and no effect on levels of
significance when analyses were performed.
40
constructs of significance" (Parker, Tupling and Brown, 1979, p.l). The
subscales of the PBI (caring and over-protection) were developed from factor
analytic studies which suggested that "the behavioral and attitudinal
components of parental bonding could be reduced to these two global
dimensions of child rearing" (Burbash, Kashani and Rosenberg, 1989, p.
418). Raw scores for each scale can be used directly or scoring can yield
five categories of parental bonding (Parker, 1989). These categories are. (1)
average (defined statistically), (2) high care/low overprotection
(conceptualized as optimal bonding), (3) low care/low overprotection
(conceptualized as low or weak bonding), (4) high care/high overprotection
(conceptualized as affectionate constraint), and (5) low care/high
overprotection (conceptualized as affectionless control) (Parker, Tupling and
Brown, 1979). For the "mother-caring" and "father-caring" scales of the PBI
a higher score indicates endorsement of caring attributes for the parent (e.g.,
friendly, helpful, warm). High scores on the "mother-overprotective" and
"father-overprotective" scales indicate endorsement of attributes of parents
having generally to do with control and intrusion (e.g., didn’t let me do
things, didn’t want me to grow up).
Initial research using the PBI focused on the relationship between the
retrospective account of early parent-child relationships and later (adult)
symptoms of depression (Burbash, Kashani and Rosenberg, 1989). Results in
41
early studies showed that low care and high overprotection were associated
with higher levels of adult depression (Bifulco, Brown and Harris, 1994).
However, these authors point out that results vis-a-vis control (or over
protection) vary and that the early findings relating over-protection to
depression have failed to replicate. In addition, they point out that some
studies have reported the relationship to be in the opposite direction (i.e., an
association between depression and low parental control).
A review of studies using the PBI has reported high internal
consistency (Cronbach’s alpha from .83 to .95) and test-retest reliability (e.g.,
r = .63 to .76 at 3 weeks). Similar internal consistency was found using
these data, with Cronbach’s alpha computed at .94 for the m o t h e r -c a r in g
subscale, .89 for the m o t h e r o v e r -p r o t e c t iv e subscale, .93 for the f a t h e r -
c a r in g subscale, and .8 8 for the f a t h e r o v e r -p r o t e c t iv e subscale. Validity has
been examined by comparison with other instruments and by examination of
construct validity. Published estimates of validity are somewhat lower than
levels of reliability (Parker, 1989).
The Early Sexual Experiences Questionnaire was derived from The
Child Maltreatment Interview Schedule published by John Briere (Briere,
1992), and consists of straight-forward questions about experiences with
sexual contact before the age of 17. The first section of the survey is in three
parts, each asking about a different type of sexual contact: (1) sexual kissing
42
and/or non-genital touching, (2) genital touching, and (3) penetration (oral,
anal or vaginal). In each case the respondent was asked if the event(s)
occurred with someone five or more years older. If so, the respondent was
asked to estimate the number of times this type of contact occurred, with
whom the contact occurred, and whether force or intimidation were used.
Questions in the next section of the survey concerned whether the respondent
was able to talk to someone about the experience and if so, when and
whether she felt like she was believed; whether the respondent ever talked to
a therapist about the event, and if so when; and whether the respondent thinks
she has remembered all of the experience(s) or has at some time forgotten
parts of the experience(s).
Variables Used in This Study
Variables used in this study are listed in Table 1. Variable names are printed
in small uppercase letters (e.g., a d u l t d e p r e s s iv e s y m p t o m s ). Most variables
were coded directly from information on the survey form (e.g. "Would you
say that you were physically abused as a child?"). Many of the questions on
the survey form provided "yes," "no," or "uncertain" as alternatives.
Analyses were performed using these variables with "uncertain" coded first as
"yes" and then with "uncertain" treated as a missing value. There were
differences in the value of the statistic, however, no differences in
43
Table 1. Variables Used in the Study
DEMOGRAPHIC VARIABLES
AGE (years)
EDUCATION (years)
INCOME
WHERE DOES RESPONDENT LIVE?
RESPONDENT’S ETHNICITY
SEXUAL ORIENTATION
GENERALLY OUT?
PART OF THE LESBIAN COMMUNITY?
FAMILY ENVIRONMENT VARIABLES
DID A PARENT HAVE PSYCHOLOGICAL TREATMENT?
DID EITHER PARENT HAVE A PSYCHIATRIC HOSPITALIZATION?
DID EITHER PARENT DRINK ALCOHOL TO EXCESS?
HOW WOULD YOU RATE YOUR CHILDHOOD?
SEXUAL ABUSE VARIABLES - OCCURRENCE
CSA (Did sexual abuse occur according to study definition?)
DID YOU EXPERIENCE SEXUAL ABUSE?
DID YOU EXPERIENCE PHYSICAL ABUSE?
WAS THERE EITHER SEXUAL OR PHYSICAL ABUSE REPORTED?
SEXUAL ABUSE VARIABLES - CIRCUMSTANCES/SEVERITY
PERPETRATOR (Who was the perpetrator o f sexual abuse?)
HOW MANY TIMES DID SEXUAL ABUSE OCCUR?
WHAT WAS THE NATURE OF THE SEXUAL ABUSE?
WAS FORCE USED IN CONJUNCTION WITH SEXUAL ABUSE?
WAS INTIMIDATION USED IN CONJUNCTION WITH SEXUAL ABUSE?
SEXUAL ABUSE VARIABLES - HELP AND SUPPORT
DID YOU TALK ABOUT YOUR SEXUAL EXPERIENCE AT THE TIME?
COULD YOU TALK ABOUT THE SEXUAL EXPERIENCE AND FEEL BELIEVED?
HAVE YOU EVER TALKED ABOUT THE SEXUAL EXPERIENCE?
HOW LONG WAS IT UNTIL YOU COULD TALK ABOUT THE SEXUAL
EXPERIENCE?
DID YOU TALK TO A THERAPIST?
HOW LONG BEFORE YOU TALKED TO A THERAPIST?
SURVEY MEASURES
ADULT DEPRESSIVE SYMPTOMS (CES-D score)
SELF-ESTEEM (Rosenberg SEI score)
MOTHER-CARING (PBI subscale)
MOTHER OVER-PROTECTIVE (PBI subscale)
FATHER-CARING (PBI subscale)
FATHER OVER-PROTECTIVE (PBI subscale)
44
significance (p) between the two sets of analyses. For this study, all analyses
were performed with "uncertain" treated as a missing value. This approach
was judged to be the more conservative, with the "uncertain" responses left
out of both the group reporting sexual abuse and the group not reporting
sexual abuse.
Other variables were computed from data provided on survey forms.
Computation of these variables is described below.
CSA
This variable was coded from the Early Sexual Experiences Questionnaire
and answers the question: Did sexual abuse occur according to study
definition?. It was coded positive (yes=l) if a respondent indicated that
before the age of 17 she had been either kissed or touched non-genitally in a
sexual way, involved in genital touching, or involved in acts of penetration
(oral, anal or vaginal) and indicated that at least one of those experiences was
with someone five or more years older than the respondent. Otherwise it was
coded negative (no=0). This definition of childhood sexual abuse was used
in all analyses and discussion unless otherwise specified. Respondents were
also asked whether they consider themselves to have been sexually abused as
children ( d id y o u e x p e r ie n c e s e x u a l a b u s e?) or physically abused as children
45
( d id y o u e x p e r ie n c e p h y s ic a l a b u s e ?), and could answer "yes," "no," or
"uncertain."
WAS THERE EITHER SEXUAL OR PHYSICAL ABUSE REPORTED?
This variable was coded "no" (0) if the respondent answered no to both the
question about whether she had experienced sexual abuse as a child and
whether she had experienced physical abuse as a child. It was coded "yes"
(1) if either of these questions was answered "yes."
PERPETRATOR
This variable answers the question: Who was the perpetrator o f the abuse?
It was coded in the following manner: (0) no abuse, (1) the abuser was not a
member of the respondent’s family, (2) the abuser was a member of the
respondent’s extended family (e.g., uncle, grandfather), (3) the abuser was a
member of the respondent’s immediate family (mother, father, step-father, or
sibling). In cases where there was more than one perpetrator (a common
occurrence) the one most closely related to the respondent was used in the
coding. For example, if a respondent indicated that she had been abused by a
stranger, her uncle and her brother, this variable was coded as #3 (immediate
family).
46
HOW MANY TIMES DID SEXUAL ABUSE OCCUR?
Some respondents replied to this question by filling in a number, and these
were coded directly into a "frequency of abuse" variable for each of the three
types of sexual experience queried in the survey. Some others entered
categorical responses such as "too often to count" or "a few times." These
were handled in two different ways. All numeric answers were coded as they
were given, up to 60. Answers greater than 60 were coded as 61. In all
cases the response "a couple" was coded as 2, "a few" as 3, "several" as 4,
"more than 10" as 15, "more than 1" as 2, and "too many to count" as 61.
Two sets of variables were coded as above, and with the following
differences: in one set the answer "I don’t know" was coded as a missing
value, and in the second that answer was coded as the mean of all given
numeric answers (8). This involved 39 respondents for the first question
(kissing and non-genital touching), 30 for the second (genital touching) and
37 for the third question (penetration). The implication of the "I don’t know"
answer seemed (based on a number of comments written by participants in
the margin of the form) to be more that it happened too many times for the
respondent to remember the number, than that the respondent had no memory
of the frequency. The variable with "don’t know" coded as the mean (8)
was, therefore, used in analyses. However, analyses were also run with the
"don’t know=missing" variable, with no difference in level of significance.
47
The question of frequency of sexual contact was asked, and a variable
coded, for each of the three sexual experience questions (non-genital
touching, touching, and penetration). A composite variable, t o t a l n u m b e r o f
a b u s e e v e n t s , was then computed by adding the responses to the three
individual questions.
WHAT WAS THE NATURE OF THE SEXUAL ABUSE?
This variable was coded with (0) meaning no abuse, (1) meaning sexual
kissing or (non-genital) touching, (2) genital touching, and (3) meaning
penetration. For respondents who experienced more than one kind o f abuse,
the most intrusive abuse (penetration>genital touching>non-genital touching)
was coded.
WAS FORCE USED IN CONJUNCTION WITH SEXUAL ABUSE? a n d W AS INTIMffiATION USED IN
CONJUNCTION WITH SEXUAL ABUSE?
These variables were coded as positive if responses to any of the three
questions (touching, genital touching or penetration) indicated that force
and/or intimidation were used during the sexual experience.
48
HOW LONG WAS IT UNTIL YOU COULD TALK ABOUT THE ABUSE? a n d HOW LONG BEFORE
YOU TALKED TO A THERAPIST?
Some respondents answered these questions by filling in a number of days,
months or years. Others responded more categorically by saying things like
"right away" or "years later." These responses were all coded into the
following categories: (1) immediately (immediately, same day, hours); (2) a
few days to a week; (3) a few weeks to a month (includes "soon"); (4) two
to twelve months; (5) one to five years; (6) six to ten years; (7) more than
ten years.
CES-D Score ( a d u l t d e p r e s s iv e s y m p t o m s ) . Rosenberg-SEI Score ( s e l f -e s t e e m s
and PBI Scores
Standard scoring was used for these instruments. First, questions were
converted so they all could be scored in the same direction. Then, for the
CES-D and Rosenberg-SEI, scores were simply added. Scores for the
subscales of the PBI were added separately for each of two parental figures,
yielding four variables, m o t h e r -c a r in g , m o t h e r o v e r -p r o t e c t iv e , f a t h e r -
c a r i n g , and f a t h e r o v e r -p r o t e c t iv e . CES-D scores were compared to the
published cutoff for depressive symptoms, and all scores were compared to
their group means.
49
Significance Levels
Significance levels of .05 are designated by a single asterisk (e.g., .123*),
levels of .01 by a double asterisk (e.g., .189**), and levels of .001 by bold
face type and a double asterisk (e.g., .234**). Because of the relatively large
number of statistical tests performed in analyzing the data, significance levels
less than .001 are to be viewed with caution. For example, if a large number
of correlations are reported, there is a virtual certainty that some are
significant by chance. For example, if 2 correlations are reported at the .05
level, there is a .10 likelihood that one of them is significant by chance, and
if 20 correlations are reported, there would be a 1.0 likelihood that one of
them is significant by chance, and so on.
Analyses
The following methods were used to analyze data in conjunction with the
four research questions posed by this study.
(1) Comparisons of a d u l t d e p r e s s iv e s y m p t o m s in CSA and non-CSA groups.
Depressive symptoms in CSA and non-CSA groups were compared by means
of analysis of variance (ANOVA).
50
(2) Analysis of variables as mediators. A path analysis model was designed
using the variables s e l f -e s t e e m , m o t h e r -c a r in g , m o t h e r o v e r -p r o t e c t iv e ,
f a t h e r -c a r in g , and f a t h e r o v e r -p r o t e c t iv e as the intervening variables
(potential mediators). The relationships between variables in the model are
shown in Figure 2. This model presumes causality, with the variables further
to the left influencing those to their right. (Biddle and Marlin, 1987; Kenny,
1979). In this analysis simultaneous regression was used, in which the CES-
D score ( a d u l t d e p r e s s iv e s y m p t o m s ) was the dependent variable. Path
coefficients were examined to determine significance of paths. In cases
where a path was seen to be significant, the indirect effects were determined
as described by Cohen and Cohen (1983).
Because this path analysis included the s e l f -e s t e e m variable, analysis
was performed using both the full CES-D score ( a d u l t d e p r e s s iv e s y m p t o m s )
and the shortened version of the CES-D ( c e s -d a b b r e v ia t e d ).
(3) Analysis of moderators. If a variable has a moderating effect, it is
expected that the different levels of the variable will produce different
relations between the predictor and the dependent variable. A moderator acts
in such a way that the covariation between two variables differs depending on
the level of the moderator (Baron and Kenny, 1986). To examine these
relationships all variable values were centered to remove non-essential
51
Figure 2. Model proposed for path analysis. Model proposes that c s a is the
independent variable, s e l f -e s t e e m and the variables derived from
the PBI ( m o t h e r -c a r in g , m o t h e r o v e r -p r o t e c t iv e , f a t h e r -c a r in g ,
and f a t h e r o v e r -p r o t e c t iv e ) are postulated to be intervening
variabels (i.e., potential mediators). The dependent variable is
a d u l t d e p r e s s iv e s y m p t o m s and is shown to the left. Positive
relationships are drawn with solid lines, and negative relationships
are drawn with a dashed line.
SELF-ESTEEM
M OTH ER-CA RING
IOTHER
IVER-PROTECTIVE.
ADULT
DEPRESSIVE
SYM PTOM S
CSA
FATHER-CARING
TA TH ER
PV ER -PR O TEC TIV ]
52
multicollinearity (Aiken and West, 1991) by converting to z-scores. An interaction
variable was computed by multiplying the c s a variable by the potential moderator
(e.g., c s a * m o t h e r -c a r in g ). Regression equations were then constructed for each
potential moderator. Each regression equation included as independent variables
the potential moderator, c s a and the interaction variable, which were entered into
the equation in that order. Significance of the interaction variable would indicate
that the potential moderator was indeed acting as a moderator.
53
RESULTS
Survey Returns
Of the total of 929 forms distributed, 493 were completed and returned, giving a
response rate of 53%. This return rate is, no doubt, a high estimate of the actual
return rate because there was no way to tally how many people refused to
accept forms at "coming out day," how many may have looked at a form in a
bookstore, but not taken it with them, or how many people may have xeroxed
the form and given it to friends.
Geographically, forms were returned from 31 states, Washington, DC,
Canada, U.S. military FPO, and Puerto Rico. The largest number of returns
came from California, with a total of 228 forms (46.2%). Regionally, the rest of
the forms were distributed as follows: Northeast (New England + New York),
118 (23.9%); Midwest, 42 (8.5%); West/Southwest, 31 (6.3%); Mid-Atlantic, 28
(5.7%); Pacific Northwest, 24 (4.9%); South, 14 (2.8%); Alaska, 5 (1.0%);
Other (U.S. military FPO, Canada, Puerto Rico), 3 (0.7%). Seven (1.4%) of the
returned forms were from women currently in correctional institutions.
54
Descriptive Data
Demographic Data
Two participants left the page of questions about demographics blank. The
following sections describe the answers of the 491 who did answer the
demographic questions.
AGE
The age range of respondents was from 16 to 88, with a mean age of 35.2 years.
Age was distributed fairly evenly over the decades, with an under-representation
of older lesbians (See Table 2 for a break-down of age, ethnicity, income and
education, as well as comparisons to (a) the recent NIMH-sponsored national
lesbian health care study, and (b) to general population statistics.)
Most participants (82.4%) described themselves as White, Caucasian or
European-American. The next largest group was Latina (5.1%). Remaining
ethnic identifications were: African-American, 4.9%; Asian or Pacific Islander,
2.2%; Native American, 1.2% and other identifications, 0.6%. There was no
significant correlation between a g e and any of the survey variables (CES-D,
Rosenberg-SEI or PBI) with one exception: There was a significant correlation
between a g e and the m o t h e r -c a r in g subscale of the PBI (r = .237; p < .000)
suggesting an association between the age of the respondent and how caring her
mother is perceived or remembered to be. There was no significant correlation
between a g e and c s a .
Table 2. Comparison of study demographic data with data from the
National Lesbian Healthcare Study and with U.S. Census Data
v a r ia b l e CURRENT
STUDY
N=493
BRADFORD, RYAN
AND ROTHBLUM
(1994)
N=],917
U.S. CENSUS
DATA ON THE
ADULT WOMAN
IN 1980 1
AGE
17-24 11.7 8.8 16.8
25-34 39.8 48.0 22.4
35-44 30.8 32.2 16.2
45-54 13.6 7.0 13.1
55 and older 3.0 3.1 31.5
98.92 99.1 100.02’
ETHNIC IDENTIFICATION
White/European 82.4 88.2 83.1
Black/African American 4.9 5.6 11.7
Latina 5.1 4.2 6.4
Asian/Pacific Islander 2.2 0.8 1.2
Native American 1.2 0.6 0.6
Other 0.6 0.3 3.0
96.4* 99.7 106.0
INCOME
9.9K or less 17.8 27.6 —
10.OK - 19.9K 17.2 35.8
20.OK - 29.9K 20.7 23.5
30.0K - 39.9K 16.6 7.9
40.0K or more 25.3 4.2
97.6 99.0
EDUCATION
Less than high school
(0-11 years) 2.0 2.4 29.1
High school graduate 9.5 9.5 37.9
Some college
or post-high school
training 25.0 18.8 15.3+5
College graduate 25.4 26.2 not available
Graduate studies
or degree 38.0 42.8 17.7+6
99.9 99.7 N/A
Quoted in Bradford, Ryan and Rothblum, 1994.
0.2% under 17 (1 16-year old responded) + 1.0% did not answer this question.
Based on adult female population 17 years old or over.
3.6% o f participants did not answer this question.
Census data does not include vocational training
Graduate degree statistic not available.
56
EDUCATION and INCOME
Most respondents had a great deal of education, with 16 years (college
graduate) being the mean, median and modal response. The range of years o f
education extended from 0 years of formal education reported by one
respondent to 24 years of formal education reported by 6 respondents. Two
percent of respondents did not graduate from high school, and 9.5%
graduated from high school, but went no further. In spite of considerable
education, incomes were low with 17.8% reporting earnings less than $10,000
per year. Results for this particular income bracket may be somewhat
misleading because of the likelihood that a number of respondents reporting
income at this level were still attending graduate school. However, income
brackets of $10K to 19.9K (17.2%), 20K to 29.9K (20.7%) and 30K to 39.9K
(16.6%) described over 50% of the sample. The percent of respondents in
each of the remaining income brackets are shown in Table 2. In summary,
63.2% of respondents had at least one college degree, while 72.3% earned
less than $40,000 last year.
There was a significant correlation between in c o m e and a d u l t
d e p r e s s iv e s y m p t o m s (r = -.203; p < .001) and between in c o m e and se l f-e s t e e m
(r = -.239, p < .001). This finding suggests that there is an association
between more or more severe depressive symptoms in adulthood and lower
57
income, and between low self-esteem and lower income. There was no
significant correlation between in c o m e and/or e d u c a t io n and history of CSA.
SEXUAI. ORIENTATION a n d PART OF THE LESBIAN COMMUNITY
Most participants identified themselves as being lesbian (89.5%), however
some said they were bisexual (8.7%), heterosexual (0.4%), asexual (0.2%),
and other (0.8%). Those checking "other" wrote in "sometimes lesbian and
sometimes heterosexual" (0.4%), and "don’t use labels" (0.4%). The
majority o f respondents (78.5%) said that they were generally out, while
20.5% said that they w ere not. A few (3) respondents said that they were not
sure how to answer the question, and were grouped with "missing data." The
majority (63.7%) also said that they were active members of the lesbian
community, however, a substantial number (34.1%) said that they did not
consider themselves to be members o f the lesbian community. The
remaining 2.2% were uncertain (5 respondents) or did not answer the
question (6 respondents)- There were no significant correlations between
variables describing sexual orientation and involvement in the lesbian
community and any of the survey scores (CES-D, Rosenberg-SEI, or PBI) or
a history o f CSA.
58
Definition and Frequency of Childhood Sexual Abuse
Respondents were asked directly whether they considered themselves to have
been sexually abused as children. Of the 490 women who answered this
question, 49.7% said "no," 41.0% said "yes," and 8.7% said that they were
"uncertain."
As previously stated, a "study definition" of CSA (represented by the
variable, c s a ) was formulated and respondents were grouped into sexually
abused and not sexually abused groups based on that definition, with 41.2%
of respondents falling into the "not sexually abused" group and 57.2%
(N=282) falling into the "sexually abused" group. The remaining 1.6% were
women who either did not answer the questions or who responded with
"uncertain" to the pertinent questions, and were coded as missing values.
Using this study definition 16,2% of the respondents were grouped into the
sexually abused category, even though they did not evaluate their experiences
as abusive. In one case a woman wrote in the margin that she had been
married at age 16 to someone who was 27, and that this was "normal in the
culture" in which she lived.
Table 3 shows the numbers of respondents who agreed and disagreed
with the study’s judgement about whether they were sexually abused as
59
children. Of the respondents, 192 (38.9%) stated that they thought they had
been sexually abused as children and were also judged by the criteria of this
study to have been sexually abused as children. An additional 179 (36.3%)
Table 3. Numbers of participants who reported having experienced
CSA and numbers of participants who were judged by
study criteria to have been sexually abused as children.
Presence of CSA as perceived by the respondent
No Yes
Study defined
childhood
sexual abuse
(c s a )
No 179 (36.3% ) 8 ( 1.6 %)
Yes 6 6 (13.4% ) 192 (38.9%)
M issing cases = 48 (9.7% )
60
stated that they were not sexually abused as children and were also judged
not to have experienced CSA by the criteria of the study. In 66 cases
(13.4%) respondents stated that they did not think they were sexually abused
as children while according to the criteria of the study they were judged to
have been sexually abused. In another 8 cases (1.6%) respondents stated that
they were sexually abused as children while the criteria of the study placed
them in the not-sexually-abused group. In these eight cases, the respondent
stated that she had at least one sexual experience before the age of 17, and
that this experience was with a person who was less than 5 years older,
therefore she did not meet the study criteria for CSA (a sexual experience
before the age of 17 with someone who was 5 or more years older).
However, these women did evaluate their early sexual experience as abusive.
In looking at these respondent’s answers to other questions in the survey, it
seems likely that she judged her early sexual experience to be abusive
because o f the nature o f the sexual experience. For example, in three cases
the sexual experience involved a family member and in four cases it involved
force or coercion.
An ANOVA was performed in order to compare the levels of the
a d u l t d e p r e s s iv e s y m p t o m s variable for each of the groups shown in Table 3.
Results of the ANOVA were significant (F = 5.90; p = .001), indicating that
there is some significant difference between a d u l t d e p r e s s iv e s y m p t o m s among
61
at least two of the groups shown in Table 3. Follow-up ANOVAs were
performed taking different groups two at a time to determine which of the
four groups shown in Table 3 differ significantly in a d u l t d e p r e s s iv e
s y m p t o m s . Results are shown in Table 4, and show a significant difference (p
< . 0 0 1 ) for a d u l t d e p r e s s iv e s y m p t o m s only between groups who were in
agreement with the judgement based on the study criteria that they did or did
not experience sexual abuse as children.
Similar follow-up ANOVAs were performed to determine which of
the groups of respondents shown in Table 3 differed significantly on the other
relevant study variables ( s e l f -e s t e e m , and the four PBI sub-scales). Those
results are shown in Tables 5 through Table 9. In all cases a significant
difference in the level of the tested variable ( a d u l t d e p r e s s iv e s y m p t o m s , s e l f
e s t e e m , and the four PBI sub-scales) was found between the two groups
agreeing with the study determination that they had or had not experienced
sexual abuse in childhood.
Those respondents who reported that they were not sexually abused as
children, while based on the criteria used in this study they were judged to
have experienced CSA (N = 64), saw their mothers and fathers as less caring
than those who agreed with the determination based on the study criteria that
they had experienced CSA (p < .001). In addition, those who were judged
by the study criteria to have experienced CSA but who reported that they
62
Table 4. Mean differences in a d u l t d e p r e s s iv e s y m p t o m s for respondents
who agree and who disagree with the study judgement about
whether they were sexually abused as children. Significance at
the .05 level is indicated by a single asterisk; at the .01 level a
double asterisk; and at the .001 level by a double asterisk and
bold face type.
Evaluation
by study
definition
Respondent’s
perception
o f CSA
N
Mean
ADULT
DEPRESSIVE
SYMPTOMS
F
P
no no 176 17.69
no yes 8 21.38 1.017 .315
no yes 8 21.38
yes no 64 19.91 0.126 .723
yes no 64 19.91
yes yes 184 22.57 2.465 .118
yes no 64 19.91
no no 176 17.69 2.235 .136
yes yes 184 22.57
no no 176 17.69 17.548 .000**
yes yes 184 22.57
no yes 8 21.38 .075 .785
63
Table 5. Mean differences in s e l f -e s t e e m for respondents who agree and
who disagree with the study judgement about whether they were
sexually abused as children. Significance at the .05 level is
indicated by a single asterisk; at the .01 level a double asterisk;
and at the .001 level by a double asterisk and bold face type.
Evaluation
by study
definition
Respondent’s
perception
of CSA
N
Mean
SELF-
ESTEEM1
F
P
no no 178 15.90
no yes 8 17.13 0.465 .496
no yes 8 17.13
yes no 64 17.44 0.025 .875
yes no 64 17.44
yes yes 190 18.28 1.091 .297
yes no 64 17.44
no no 178 15.90 4.321 .039*
yes yes 190 18.28
no no 178 15.90 18.218 .0 0 0 **
yes yes 190 18.28
no yes 8 17.13 .323 .570
The Rosenberg-SEI was scored such that higher scores represent lower self-esteem.
64
Table 6 . Mean differences in m o t h e r -c a r in g for respondents who agree and
who disagree with the study judgement about whether they were
sexually abused as children. Significance at the .05 level is
indicated by a single asterisk; at the .01 level a double asterisk;
and at the .001 level by a double asterisk and bold face type.
Evaluation
by study
definition
Respondent’s
perception
o f CSA
N
Mean PBI
MOTHER-
CARING
F
P
no no 167 13.57
no yes 8 14.38 0.056 .814
no yes 8 14.38
yes no 62 15.26 0.062 .804
yes no 62 15.26
yes yes 183 20.19 13.011 .0 0 0 **
yes no 62 15.26
no no 167 13.57 1.506 .2 2 1
yes yes 183 20.19
no no 167 13.57 43.861 .0 0 0 **
yes yes 183 20.19
no yes 8 14.38 2.885 .091
65
Table 7. Mean differences in m o t h e r o v e r -p r o t e c t iv e for respondents who
agree and who disagree with the study judgement about whether
they were sexually abused as children. Significance at the .05
level is indicated by a single asterisk; at the .01 level a double
asterisk; and at the .001 level by a double asterisk and bold face
type.
Evaluation
by study
definition
Respondent’s
perception
of CSA
N
Mean PBI
MOTHER
OVER-
PROTECTIVE
F
P
no no 169 25.09
no yes 8 23.25 0.352 .554
no yes 8 23.25
yes no 64 22.27 0.082 .776
yes no 64 22.27
yes yes 183 20.51 1.764 .185
yes no 64 22.27
no no 169 25.09 4.746 .030*
yes yes 183 20.51
no no 169 25.09 23.698 .0 0 0 **
yes yes 183 20.51
no yes 8 23.25 0.721 .397
66
Table 8 . Mean differences in f a t h e r -c a r in g for respondents who agree and
who disagree with the study judgement about whether they were
sexually abused as children. Significance at the .05 level is
indicated by a single asterisk; at the .01 level a double asterisk;
and at the .001 level by a double asterisk and bold face type.
Evaluation
by study
definition
Respondent’s
perception
o f CSA
N
Mean PBI
FATHER-
CARING
F
P
no no 162 16.94
no yes 6 16.33 0 . 0 2 2 .883
no yes 6 16.33
yes no 54 17.44 0.084 .772
yes no 54 17.44
yes yes 170 22.35 11.249 .0 0 1 **
yes no 54 17.44
no no 162 16.94 0.109 .742
yes yes 170 22.35
no no 162 16.94 25.423 .0 0 0 **
yes yes 170 22.35
no yes 6 16.33 2.267 .134
67
Table 9. Mean differences in f a t h e r o v e r -p r o t e c t iv e for respondents who
agree and who disagree with the study judgement about whether
they were sexually abused as children. Significance at the .05
level is indicated by a single asterisk; at the .01 level a double
asterisk; and at the .001 level by a double asterisk and bold face
type.
Evaluation
by study
definition
Respondent’s
perception
of CSA
N
Mean PBI
FATHER
OVER-
PROTECTIVE
F
P
no no 159 27.70
no yes 6 23.67 1.703 .194
no yes 6 23.67
yes no 57 23.02 0.037 .848
yes no 57 23.02
yes yes 164 22.17 0.398 .529
yes no 57 23.02
no no 159 27.70 16.450 .0 0 0 **
yes yes 164 22.17
no no 159 27.70 36.076 .0 0 0 **
yes yes 164 22.17
no yes 6 23.67 0.158 .691
68
were not sexually abused as children had lower self-esteem (p < .05), saw
their fathers (p < .001) and mothers (p < .05) as less "over protective," than
those who agreed with the study determination that they had not experienced
CSA.
No significant differences were found between the group of
respondents who disagreed with the study determination that they were not
abused (by judging themselves to have been sexually abused) and other
groups. However, it must be noted that this group of respondents was so
small (N=8) that there may be insufficient power for comparisons to reach
significance.
In summary, for those groups of respondents who agreed with the
study determination about their abuse status, significant differences were
found between for all variables tested between those who were judged to
have experienced and not experienced CSA. The group of respondents who
disagreed with the study determination that they were sexually abused as
children by stating that they did not consider themselves to have been abused,
saw their parents as significantly less caring than those who were in
agreement with the study determination that they were abused. Other
significant differences are consistent with findings for respondents who
agreed with the study determination of their CSA status. Finally,
comparisons for respondents who reported CSA, but were judged by study
69
criteria to not have been abused were not significant, but may not have had
sufficient power to reach significance.
Comparison of CSA Group with Non-CSA Group
The CSA and non-CSA groups, as defined by study criteria, were compared
on all demographic variables. Results are shown in Table 10, and
demonstrate no differences significant to the .001. The a g e variable was
significant at the .05 level, with the CSA group of respondents being
somewhat older (35.9 years v. 34.1 years for the non-CSA group). The
e d u c a t io n variable was significant at the .01 level, with those in the CSA-
group having somewhat less education (mean = 15.74 years) than non-CSA
respondents (mean = 16.53 years). These levels of significance may be
viewed as a possible trend.
Correlational Data. Pearson correlations for all variables (excluding
demographic variables) are presented in Tables 11 to 13. These tables show
correlations for all respondents (Table 11), for respondents judged not to have
been sexually abused as children (Table 12), and respondents judged to have
experienced CSA (Table 13). Important correlations will be discussed in the
appropriate sections below.
70
Table 10. Comparison of respondents who reported sexual abuse with
respondents who did not report sexual abuse (Using ANOVA).
Significance at the .05 level is indicated by a single asterisk; at
the .01 level a double asterisk; and at the .001 level by a
double asterisk and bold face type.
Variable Group N Mean F
P
AGE not abused 2 0 1 34.1
abused 281 35.9 4.15 .042*
INCOME not abused 199 3.65
abused 276 3.34 2.47 .117
EDUCATION not abused 2 0 2 16.53
abused 282 15.74 9.38 .0 0 2 **
SEXUAL not abused 2 0 2 1.09 1
ORIENTATION abused 281 1.17 2 .2 1 .138
GENERALLY not abused 2 0 2 0.78 2
OUT abused 281 0.80 0.67 .415
ACTIVE
MEMBER OF not abused 2 0 1 0 . 6 8 '
LESBIAN abused 278 0.62 0.46 .500
COMMUNITY
Categorical coding: (l)=lesbian, (2)=bisexual, (3)=heterosexual, (4)=asexual, (5) other ("sometimes
lesbian, sometimes heterosexual" or "don't use labels." Means further from 1.00 indicate greater
divergence from a purely lesbian identification.
Categorical variable with coding: (0)=no; (1 )=yes. Means further from zero indicate more
"outness," or more involvement with the lesbian community.
71
Table 11. Correlation matrix of variables (excluding demographics) for
all participants. This table contains two-tailed correlations
between variables (as listed in Table 1 and excluding
demographic variables). All participants are included. A
single asterisk marks correlations significant at the .05 level; a
double asterisk marks significance at the .01 level; and a
double asterisk and bold face type mark correlations significant
to the .001 level. A key relating the variable names to
numbers used in the matrix is given at the bottom of the page.
1 2 3 4
5
6 7 8 9 10
1 1.000
2 .576** 1.000
3 .185** .189** 1.000
4 -.253** -.184** -.213** 1.000
5 .072 .074 .192** - .258**
1.000
6 .080 .046 .251** -.356**
.696**
1.000
7 .131** .136** .345** -.443**
294**
.428** 1.000
8 .047 .048 .097* -.070
• 377**
.343** .292** 1.000
9 .063 .049 .192** -.301**
.776**
.650** .324** .554** 1.000
10 .007 .062 .302** -.354**
.753**
.695**
3 5 7 **
.192* .597** 1.000
11 .063 .026 .157** - .248**
.894**
.712** .315** .560** .831** .680**
12 .017 .030 .125* -.234**
.508**
.589** .401** .096* .338** .735**
13 .027 .058 .191** -.237**
• 627**
.689** 292** .117* .416** .708**
14 .005 .032 -.002 .124**
-049
-.103* -.089 -.022 -.125** .011
15 -.030 .018 .035 .162**
-120*
-.196** -.150** -.092 -.192** -.074
16 .138** .062 .175** -.134**
• 331**
.349** . 1 0 2 * .145** .239** .324**
1 = DID A PARENT H A V E PSYCHOLOGICAL TREATMENT?
2 = DID EITHER PARENT HAVE A PSYCHIATRIC HOSPITALIZATION7
3 = DID EITHER PARENT DRINK ALCOHOL TO EXCESS?
4 = HOW WOULD Y O U RATE YOUR CHILDHOOD?
5 = CSA (Did sexual abuse occur according to study definition?)
6 = DID YOU EXPERIENCE SEXUAL ABUSE?
7 = DID YOU EXPERIENCE PHYSICAL ABUSE?
8 = WAS THERE EITHER SEXUAL OR PHYSICAL ABUSE REPORTED?
9 = PERPETRATOR
10 = HOW MANY TIM ES DID SEXUAL ABUSE OCCUR?
11 = WHAT WAS THE NATURE OF THE SEXUAL ABUSE?
12 = WAS FORCE USED IN CONJUNCTION W ITH SEXUAL ABUSE?
13 = WAS INTIMIDATION USED IN CONJUNCTION WITH SEXUAL ABUSE?
14 = DID YOU TALK A BOU T YOUR SEXUAL EXPERIENCE AT iH E TIME?
15 = COULD YOU TA LK ABOUT THE SEXUAL EXPERIENCE AND PEEL BELIEVED?
16 = HAVE YOU EVER TALKED ABOUT TH E SEXUAL EXPERIENCE?
72
Table 11 (Continued).
Correlation matrix of variables (excluding demographics) for all
participants. Two-tailed correlations were run between all variables
used in this study (as listed in Table 1 and excluding demographic
variables). All participants are included. A single asterisk marks
correlations significant at the .05 level; a double asterisk marks
significance at the .01 level; and a double asterisk and b o ld f a c e type
mark correlations significant to the .001 level. A key relating the
variable names to numbers used in the matrix is given at the bottom of
the page.
1 2 3 4 5 6 7 8 9 10
1 7 -.141** -.040 -.114* .138** -.342** -.391** -.107* -.122* -.233** -.319**
18 .114* .039 . 2 0 0 ** -.319** .423** .533** .276** .239** .410** .517**
1 9 -.088 -.015 -.153** .281** -.411** -.524** -.240** - . 2 0 1 ** -.376** -.490**
20 .111* .044 .093* -.296** .159** .185** .180** .086 .162** .256**
21 .108* .079 .111* -.339** .179** .174** .251** .050 .154** .308**
22 .127** .118* .288** -.607** .247** .300** .404** .121** .284** .340**
23 -.066 -.018 -.082 .359** -.187** - . 2 0 0 ** -.205** -.105* -.188** -.227**
24 .163** .136** .191**
..545**
.195** .251** .434** .125** .229** .296**
25 -.052 -.066 -.143** .339** -.269** -.245** -.335** -.130** -.232** -.362**
1 = DID A PARENT HAVE PSYCHOLOGICAL TREATMENT?
2 = DID EITHER PARENT HAVE A PSYCHIATRIC HOSPITALIZATION?
3 = DID EITHER PARENT DRINK ALCOHOL TO EXCESS?
4 = HOW WOULD YOU RATE YOUR CHILDHOOD?
5 = CSA (Did sexual abuse occur according to study definition?)
6 = DID YOU EXPERIENCE SEXUAL ABUSE?
7 - DID YOU EXPERIENCE PHYSICAL ABUSE?
8 = WAS THERE EITHER SEXUAL OR PHYSICAL ABUSE REPORTED?
9 = PERPETRATOR
10 - HOW MANY TIMES DID SEXUAL ABUSE OCCUR?
17 = HOW LONG WAS IT UNTIL YOU COULD TALK ABOUT THE SEXUAL EXPERIENCE?
18 = DID YOU TALK TO A THERAPIST?
19 = HOW LONG BEFORE YOU TALKED TO A THERAPIST?
20 = ADULT DEPRESSIVE SYMPTOMS
21 = SELF-ESTEEM
22 = MOTHER-CARING (PBI subscale)
23 = MOTHER OVER-PROTECTIVE (PBI subscale)
24 = FATHER-CARING (PBI subscale)
25 = FATHER OVER-PROTECTIVE (PBI subscale)
73
Table 11 (Continued).
Correlation matrix of all variables (excluding demographics) for all
participants. Two-tailed correlations were run between all variables
used in this study (as listed in Table 1 and excluding demographic
variables). All participants are included. A single asterisk marks
correlations significant at the .05 level; a double asterisk marks
significance at the .01 level; and a double asterisk and bold face type
mark correlations significant to the .001 level. A key relating the
variable names to numbers used in the matrix is given at the bottom of
the page.
1 1 12 13 14 15 16 1 7 18
1 1 1.000
12 .462** 1.000
13 .515** .705** 1.000
14 -.045 -.079 -.092 1.000
15 -.123* ..168** -.155** .794** 1.000
16 .293** . 2 2 0 ** .277** .301** .249** 1.000
1 7 -.296** -.276** -.315** -.007 .057 -.833** 1.0000
18 .428** .365** .409** -.066 -.155** .416** -.412** 1.0000
19 -.406** -.362** -.372** .104* .188** -.396** .469** -.941**
20 .131** .171** .176** -.040 -.099* .102* -.101* .244**
21 .127** .189** . 2 0 1 ** -.006 -.069 .094 -.106* .250**
22 .263** .217** .230** -.103* -.151** .069 -.074 .312**
23 -.213** -.156** -.180** -.069 -.044 -.117* .075 -.182**
24 .185** .213** .182** -.085 -.130* .107* -.118* .268**
25 -.257** -.244** -.205** -.013 -.007 -.115* .096 -.224**
11 = WHAT WAS THE NATURE OF THE SEXUAL ABUSE?
12 = WAS FORCE USED IN CONJUNCTION WITH SEXUAL ABUSE?
13 - WAS INTIMIDATION USED IN CONJUNCTION WITH SEXUAL ABUSE?
14 = DID YOU TALK ABOUT YOUR SEXUAL EXPERIENCE AT THE TIME?
15 = COULD YOU TALK ABOUT THE SEXUAL EXPERIENCE AND FEEL BELIEVED?
16 - HAVE YOU EVER TALKED ABOUT THE SEXUAL EXPERIENCE?
17 = HOW LONG WAS IT UNTIL YOU COULD TALK ABOUT THE SEXUAL EXPERIENCE?
18 = DID YOU TALK TO A THERAPIST?
19 - HOW LONG BEFORE YOU TALKED TO A THERAPIST?
20 = ADULT DEPRESSIVE SYMPTOMS
21 = SELF-ESTEEM
22 = MOTHER-CARING (PBI subscale)
23 = MOTHER OVER-PROTECTIVE (PBI subscale)
24 = FATHER-CARING (PBI subscale)
25 = FATHER OVER-PROTECTIVE (PBI subscale)
74
Table 11 (Continued).
Correlation matrix of all variables (excluding demographics) for all
participants. Two-tailed correlations were run between all variables
used in this study (as listed in Table 1 and excluding demographic
variables). All participants are included. A single asterisk marks
correlations significant at the .05 level; a double asterisk marks
significance at the .01 level; and a double asterisk and bold face type
mark correlations significant to the .001 level. A key relating the
variable names to numbers used in the matrix is given at the bottom of
the page.
19 20 21 22 23 24 25
19 1.000
20 -.218** 1.000
2 1 ■.241** .730** 1.000
22 ..261** .310** .283** 1.000
23 .142** -.317** -.301** ■.397** 1.000
24 ..249** .196** .215** .530** -.231** 1.000
25 .231** - . 2 2 2 ** ..207** -.317** .622** -.365**
19 = HOW LONG BEFORE YOU TALKED TO A THERAPIST?
20 = ADULT DEPRESSIVE SYMPTOMS
21 = SELF-ESTEEM
22 - MOTHER-CARING (PBI subscale)
23 = MOTHER OVER-PROTECTIVE (PBI subscale)
24 = FATHER-CARING (PBI subscale)
25 = FATHER OVER-PROTECTIVE (PBI subscale)
75
Table 12. Correlation matrix of variables (excluding demographic
variables and variables concerned with the nature of abuse) for
participants judged not to have experienced CSA. All
correlations are two-tailed. Only participants judged not to
have been sexually abused as children are included. A single
asterisk marks correlations significant at the .05 level; a double
asterisk marks significance at the .01 level; and a double
asterisk and bold face type mark correlations significant to the
.001 level. A key relating the variable names to numbers used
in the matrix is given at the bottom of the page.
1 2 3 4 14 15 16 17
1 1.000
2 .518** 1.000
3 .217** .197** 1.000
4 -.324** -.081 -.172* 1.000
14 -.004 .110 .146 .026 1.000
15 -.027 .107 .166* .050 .930** 1.000
16 .095 .059 .185* -.072 .525** .484** 1.000
1 7 -.096 -.004 -.026 .047 -.184* -.148 - .877** 1.000
18 .114 .036 .134 -.157* .012 -.059 .390** ■.403**
19 -.073 .005 .000 .138 .103 .146 -.361** .483**
20 .148* -.079 .063 -.249** -.033 -.070 .053 -.086
21 .200** .053 .080 -.338** -.003 -.012 .056 -.058
22 .168* .079 .269** -.533** -.086 -.103 -.084 .093
23 -.133 -.056 -.095 .291** -.119 -. 189* -.069 .026
24 .215** .109 .156* -.589** -.040 -.054 .029 -.027
25 -.134 -.030 -.110 .213* -.137 -.153 -.058 .004
1 = DID A PARENT HAVE PSYCHOLOGICAL TREATMENT?
2 - DID EITHER PARENT HAVE A PSYCHIATRIC HOSPITALIZATION?
3 = DID EITHER PARENT DRINK ALCOHOL TO EXCESS?
4 = HOW WOULD YOU RATE YOUR CHILDHOOD?
14 = DID YOU TALK ABOUT YOUR SEXUAL EXPERIENCE A T THE TIME?
15 = COULD YOU TALK ABOUT THE SEXUAL EXPERIENCE AND FEEL BELIEVED?
16 = HAVE YOU EVER TALKED ABOUT THE SEXUAL EXPERIENCE?
17 = HOW LONG WAS IT UNTIL YOU COULD TALK ABOUT THE SEXUAL EXPERIENCE?
18 = DID YOU TALK TO A THERAPIST?
19 = HOW LONG BEFORE YOU TALKED TO A THERAPIST?
20 = ADULT DEPRESSIVE SYMPTOMS
21 = SELF-ESTEEM
22 = MOTHER-CARING (PBI subscale)
23 = MOTHER OVER-PROTECTIVE (PBI subscale)
24 = FATHER-CARING (PBI subscale)
25 = FATHER OVER-PROTECTIVE (PBI subscale)
76
Table 12. (Continued)
Correlation matrix of all variables (excluding demographic variables
and variables having to do with the nature of abuse) for participants
determined not to have experienced CSA according to the study
definition ( c s a = no). Two-tailed correlations were run between all
variables used in this study (as listed in Table 1 and excluding
demographic variables and sexual abuse variables). Only participants
judged not to have been sexually abused as children are included. A
single asterisk marks correlations significant at the .05 level; a double
asterisk marks significance at the .01 level; and a double asterisk and
bold face type mark correlations significant to the .001 level. A key
relating the variable names to numbers used in the matrix is given at
the bottom of the page.
18 19 20 2 1 22 23 24 25
18 1.000
19 -.913** 1.000
20 .116 -.105 1.000
21 .165* -.167* .709** 1.000
22 .174* -.090 .201** .214** 1.000
23 .025 -.119 -.279** -.237** -.315** 1.000
24 .199* -.137
199**
.224** .540** -.151* 1.000
25 -.070 -.003 -.185* -.099 -.196** .551** -.241** 1.000
IX = DID YOU TALK TO A THERAPIST?
19 = HOW LONG BEFORE YOU TALKED TO A THERAPIST?
20 = ADULT DEPRESSIVE SYMPTOMS
21 = SELF-ESTEEM
22 = MOTHER-CARING (PBI subscale)
23 = MOTHER OVER-PROTECTIVE (PBI subscale)
24 = FATHER-CARING (PBI subscale)
25 = FATHER OVER-PROTECTIVE (PBI subscale)
77
Table 13. Correlation matrix of all variables (excluding demographic
variables) for participants determined to have experienced CSA
according to the study definition ( c s a = yes). Two-tailed
correlations were run between all variables used in this study (as
listed in Table 1 and excluding demographic variables). Only
participants judged to have been sexually abused as children are
included. A single asterisk marks correlations significant at the
.05 level; a double asterisk marks significance at the .01 level;
and a double asterisk and bold face type mark correlations
significant to the .001 level. A key relating the variable names
to numbers used in the matrix is given at the bottom of the page.
1 2 3 4 6 7 9 10
1 1.000
2 .609** 1.000
3 .166** . 171** 1.000
4 -.177** -.234** -.183** 1.000
6 .071 -.010 .217** -.295** 1.000
7 .121 .122 .384** -.465** .361** 1.000
9 .033 .069 .181** -.324** .321** .232** 1.000
1 0 .107 .172 .112 -.107 .282** .095 .135 1.000
1 1 .014 -.016 .040 -.128* .375** .183** .135* .391**
1 2 -.042 -.015 .037 -.171** .394** .350** .141* .417**
1 3 -.053 .007 .118 -.120 .482** .173** .183** .281*
1 4 .022 -.004 -.076 .164* -.163** -.104 -.226** .214
1 5 -.021 -.028 .005 .199** -.247** -.191** -.231** -.026
1 6 .147* .038 .081 -.036 .142* .034 -.022 .055
1 = DID A PARENT HAVE PSYCHOLOGICAL TREATMENT?
2 = DID EITHER PARENT HAVE A PSYCHIATRIC HOSPITALIZATION?
3 - DID EITHER PARENT DRINK ALCOHOL TO EXCESS?
4 = HOW WOULD YOU RATE Y O U R CHILDHOOD?
5 = CSA (D id sexual abuse occur according to study definition?)
6 = DID Y O U EXPERIENCE SEXUAL ABUSE?
7 = DID Y O U EXPERIENCE PHYSICAL ABUSE?
8 = WAS THERE EITHER SEXUAL OR PHYSICAL ABUSE REPORTED?
9 = PERPETRATOR
10 = HOW M ANY TIMES DID SEXUAL ABUSE OCCUR?
11 = WHAT WAS THE NATURE OF THE SEXUAL ABUSE?
12 = WAS FORCE USED IN CONJUNCTION WITH SEXUAL ABUSE?
13 = WAS INTIMIDATION USED IN CONJUNCTION WITH SEXUAL ABUSE?
14 - DID Y O U TALK ABOUT Y O U R SEXUAL EXPERIENCE AT THE TIME?
15 = COULD YOU TALK ABOUT THE SEXUAL EXPERIENCE AND FEEL BELIEVED?
16 = HAVE Y OU EVER TALKED ABOUT THE SEXUAL EXPERIENCE?
78
Table 13 (Continued)
Correlation matrix o f all variables (excluding demographic variables)
for participants determined to have experienced CSA according to the
study definition ( c s a = yes). Two-tailed correlations were run between
all variables used in this study (as listed in Table 1 and excluding
demographic variables). Only participants judged to have been sexually
abused as children are included. A single asterisk marks correlations
significant at the .05 level; a double asterisk marks significance at the
.01 level; and a double asterisk and bold face type mark correlations
significant to the .001 level. A key relating the variable names to
numbers used in the matrix is given at the bottom of the page.
1 2 3 4 6 7 9 10
1 7 -.163* -.037 -.055 .069 -.213** -.050 -.003 -.109
18 .080 .010 .135* ■.277** .358** .223** .207** .235*
19 -.058 .008 -.108 .225** -.350** -.191** -.191** -.220
20 .064 .092 .064 ■.293** .010 .187** .128* .068
21 .029 .076 .071 -.302** .066 .238** .146* .149
22 .071 .126* .256** ■.601** .225** .405** .226** .137
23 .018 .024 -.035 .361** -.085 -.243** -.067 .161
24 .093 .136* .175** -.465** . 2 2 0 ** .435** .234** .273*
25 .042 -.056 -.099 .347** -.043 - .332** -.107 -.032
1 = DID A PARENT HAVE PSYCHOLOGICAL TREATMENT?
2 = DID EITHER PARENT HAVE A PSYCHIATRIC HOSPITALIZATION?
3 = DID EITHER PARENT DRINK ALCOHOL TO EXCESS?
4 = HOW WOULD YOU RATE YOUR CHILDHOOD?
5 = CSA (Did sexual abuse occur according to study definition?)
6 = DID YOU EXPERIENCE SEXUAL ABUSE?
7 = DID YOU EXPERIENCE PHYSICAL ABUSE?
8 = WAS THERE EITHER SEXUAL OR PHYSICAL ABUSE REPORTED?
9 = PERPETRATOR
10 = HOW MANY TIMES DID SEXUAL ABUSE OCCUR?
17 = HOW LONG WAS IT UNTIL YOU COULD TALK ABOUT THE SEXUAL EXPERIENCE?
18 = DID YOU TALK TO A THERAPIST?
19 = HOW LONG BEFORE YOU TALKED TO A THERAPIST?
20 = ADULT DEPRESSIVE SYMPTOMS
21 = SELF-ESTEEM
22 = MOTHER-CARING (PBI subscale)
23 = MOTHER OVER-PROTECTIVE (PBI subscale)
24 = FATHER-CARING (PBI subscalc)
25 = FATHER OVER-PROTECTIVE (PBI subscale)
79
Table 13 (Continued)
Correlation matrix of all variables (excluding demographic variables)
for participants determined to have experienced CSA according to the
study definition ( c s a = yes). Two-tailed correlations were run between
all variables used in this study (as listed in Table 1 and excluding
demographic variables). Only participants judged to have been sexually
abused as children are included. A single asterisk marks correlations
significant at the .05 level; a double asterisk marks significance at the
.01 level; and a double asterisk and bold face type mark correlations
significant to the .001 level. A key relating the variable names to
numbers used in the matrix is given at the bottom of the page.
11 12 13 14 1 5 16 1 7 18
11 1.000
12 .381** 1.000
13 .317** .571** 1.000
14 -.044 -.082 -.102 1.000
15 -.039 -.179** -.149* .703** 1.000
16 .019 .098 .153* .177** .130* 1.000
1 7 -.071 -.172** -. 192** .102 .167** -.759** 1.000
18 .176** .241** .248** -. 083 -.146* .274** ■.255** 1.000
19 -.196** •.239** -.198** .097 .165** -.260** .331** -.936**
20 .000 .125 .094 -.021 -.089 .056 -.046 .232**
21 .024* .130* .108 .013 -.079 .042 -.068 .231**
22 .133* .148* .133* -.088 -.137* .044 -.052 .273**
23 -.111 -.106 -.094 -.063 .017 -.057 .014 -.174**
24 .074 .180** .090 -. 095 -.152* .070 -.104 .223**
25 -.123 -.153* -.049 .028 .038 -.016 .018 -.159*
11 = WHAT WAS THE NATURE OF THE SEXUAL ABUSE?
12 = WAS FORCE USED IN CONJUNCTION WITH SEXUAL ABUSE?
13 = WAS INTIMIDATION USED IN CONJUNCTION WITH SEXUAL ABUSE?
14 = DID YOU TALK ABOUT YOUR SEXUAL EXPERIENCE AT THE TIME?
15 = COULD YOU TALK ABOUT THE SEXUAL EXPERIENCE AND FEEL BELIEVED?
16 = HAVE YOU EVER TALKED ABOUT THE SEXUAL EXPERIENCE?
17 = HOW LONG WAS IT UNTIL YOU COULD TALK ABOUT THE SEXUAL EXPERIENCE?
IS - DID YOU TALK TO A THERAPIST?
19 = HOW LONG BEFORE YOU TALKED TO A THERAPIST?
20 - ADULT DEPRESSIVE SYMPTOMS
21 = SELF-ESTEEM
22 - MOTHER-CARING (PBI subscale)
23 = MOTHER OVER-PROTECTIVE (PBI subscale)
24 = FATHER-CARING (PBI subscale)
25 = FATHER OVER-PROTECTIVE (PBI subscale)
80
Table 13 (Continued)
Correlation matrix of all variables (excluding demographic variables)
for participants determined to have experienced CSA according to the
study definition ( c s a = yes). Two-tailed correlations were run between
all variables used in this study (as listed in Table 1 and excluding
demographic variables). Only participants judged to have been sexually
abused as children are included. A single asterisk marks correlations
significant at the .05 level; a double asterisk marks significance at the
.01 level; and a double asterisk and bold face type mark correlations
significant to the .001 level. A key relating the variable names to
numbers used in the matrix is given at the bottom of the page.
19 20 2 1 22 23 24 25
19 1.000
20 -.194** 1.000
2 1 -.217** .729** 1.000
22 - . 2 2 0 ** .337** .276** 1.000
23 .151* -.297** - .303** -.406** 1.000
24 • .223** .137* .149* .480** -.224** 1.000
25 .199** -.173** -.199** -.313** .621** -.404** 1
19 = HOW LONG BEFORE YOU TALKED TO A THERAPIST?
20 = ADULT DEPRESSIVE SYMPTOMS
21 = SELF-ESTEEM
22 = MOTHER-CARING (PBI subscale)
23 = MOTHER OVER-PROTECTIVE (PBI subscale)
24 = FATHER-CARING (PBI subscale)
25 = FATHER OVER-PROTECTIVE (PBI subscale)
81
Survey Results
Descriptive statistics for the Rosenberg-SEI, CES-D, and PBI are presented in
Table 14. In addition, results are described below.
SELF-ESTEEM
The Rosenberg-SEI was completed by 487 (98.8%) of the respondents. For
those who completed this instrument, the range of scores extended from 10 to
38 out of a possible range (using reverse scoring) of 10 (high self-esteem) to
40 (low self-esteem) as shown in Table 14. The mean score was 17.22, with
the median at 16 and mode at 10, giving a profile skewed to the lower self
esteem end of the scale (Skew= 0.90).
ADULT DEPRESSIVE SYMPTOMS
The depression inventory (CES-D) was completed by 479 respondents
(97.2%). The possible range on this instrument is 0 (no symptoms) to 60 (all
symptoms strongly endorsed). Respondents in this study scored in a range
from 0 to 58 (as shown in Table 14), with a mean of 20.32, a median of 20
and a mode of 20. As previously stated, the usual cutoff for depressive
symptoms on the CES-D is 16, with those scoring above 16 viewed as having
at least moderate depressive symptoms (Devins and Orme, 1985; Hsu and
Marshall, 1987). Among respondents as a whole, 62% scored above 16 and
82
Table 14. Descriptive statistics for survey instruments (Rosenberg-SEI, CES-
D and PBI). The mean, standard deviation (SD) and range are
given for each instrument for all respondents, respondents judged
to have experienced CSA (CSA group) and the comparison group
of respondents judged not to have experienced CSA (non-CSA
group). Higer scores on the Rosenberg-SEI reflect lower self
esteem, while higher levels on the CES-D and PBI subscales
reflect higher levels o f depressive symptoms and parental caring
and over-protection respectively.
Variable Possible
Range
All
Respondents
Mean (SD) Range
CSA
Group
Mean (SD) Range
non-CSA
Group
Mean (SD) Range
SELF-ESTEEM
(Rosenberg-SEI)
10-40 17.22 (5.48) 10-38 18.08 (5.66) 10-38 16.07 (5.07) 10-37
ADULT
DEPRESSIVE
SYMPTOMS
(CES-D)
0-60 20.32 (11.35) 0-58 21.83 (11.80) 0-53 18.14 (10.32) 0-58
PBI
MOTHER-
CARING
0-36 16.57 (9.92) 0-36 18.52 (9.75) 0-36 13.57 (9.44) 0-36
PBI
MOTHER
OVER-
PROTECTIVE
0-39 22.72 (9.03) 0-39 21.43 (9.03) 0-39 24.77 (8.55) 0-39
PBI
FATHER-
CARING
0-36 19.29 (9.93) 0-36 20.88 (9.61) 0-36 17.03 (10.02) 0-36
PBI
FATHER
OVER-
PROTECTIVE
0-39 24.53 (8.50) 2-39 22.65 (8.67) 2-39 27.21 (7.54) 3-39
83
48% above 20 on this measure. For those judged to have been sexually abused
in childhood, 67% scored above 16 and 54% above 20, and for those judged not
to have experienced sexual abuse as children, 55% scored above 16 and 38%
above 20.
Parental Bonding Scores
The PBI, as stated, is composed of four subscales which were used as variables
in this study: m o t h e r -c a r in g , m o t h e r o v e r -p r o t e c t iv e , f a t h e r -c a r in g , and f a t h e r
o v e r -p r o t e c t iv e . Respondents were asked to fill out the scales based on their
earliest memories of their parental figures. If they were parented by people
other than a mother and father, they were asked to indicate who the parent
figure was. Ninety-eight percent of respondents indicated that their responses
described their mother, while 1.2% specified a stepmother (2 respondents), sister
(1 respondent), grandmother (2 respondents) or aunt (1 respondent). The
remaining 0.8% reported no mother figure or did not answer. Ninety-one
percent of respondents indicated that their responses described a father, with
4.3% (21 respondents) saying there was no father-figure in their childhood,
3.0% (15 respondents) specifying a stepfather, 0.6% (3 respondents) specifying
another woman as the second parental figure, 0.4% (2 respondents) specifying a
grandfather, and 0.8% (4 respondents) not answering the question.
84
The mean, standard deviation and range of scores on the PBI subscales
are shown in Table 14. Higher scores on the "caring" subscales indicate a
perception of the parent being more caring as evidenced by the endorsement of
attributes such as warmth, understanding, or affection; and behaviors such as
helping, smiling at the child or praising the child.
Higher scores on the "over-protective" subscales indicate a perception of
the parent being more controlling or intrusive as demonstrated by such behaviors
as not allowing the child to do what she wanted, trying to control what the child
did, and trying to make the child dependent, as well as perceived intentions such
as not wanting the child to grow up or not wanting the child to make her own
decisions.
Correlational Data and Significance of Means. Pearson correlations were
computed between PBI subscale scores and the experience of CSA ( c s a ) , and
between PBI subscale scores and a d u l t d e p r e s s iv e s y m p t o m s . These are shown in
Tables 1 1 through 13. Correlations of PBI subscales with a d u l t d e p r e s s iv e
s y m p t o m s are significant whether looking at the sample as a whole, or at either
the group that was judged to have been sexually abused as children, or the
comparison group judged not to have experienced CSA.
These correlations suggest that those who were judged to have
experienced CSA perceived their parents as more caring and less over-protective
85
that the comparison group. A further analysis was undertaken to determine
whether these perceptions of parents were specific to inter- or intra-familial
abuse. Results of this analysis are shown in Table 15, and indicate that
perception of the parent as more caring is most strongly (and significantly)
associated with sexual abuse by a first degree relative rather than by a second
degree relative or non-related person. A weaker (but significant) association
also exists between the f a t h e r -c a r in g variable and sexual abuse by a second
degree family member.
Some Parental Characteristics
A substantial number of respondents (27.2%) indicated that one or the other of
their parents had psychological treatment when the respondent was a child, 7.7%
were uncertain and 64.9% reported no psychological treatment for either parent.
In 5.3% o f all cases psychological treatment was reported for both parents, and
in 10.8% of all cases the psychological treatment of at least one parent included
hospitalization. Most respondents (76.9%) had one or more parents who drank
alcohol, and 36.1% of all respondents reported parental alcohol consumption to
be "excessive."
86
Table 15. Comparison of means of the PBI subscales for groups of
respondents reporting CSA histories with various perpetrators.
Only respondents who were judged to have experienced CSA
were used in this analysis. The first entry for each PBI subscale
contains the results of and ANOVA for all three categories of
perpetrator (non-family, second degree family, and first degree
family) If the ANOVA is significant, it is followed by ANOVAs
comparing PBI means with categories compared two at a time to
demonstrate which categories are responsible for the significance
of the original ANOVA.
Variable Non-family
mean (N)
Second-degree
family
mean (N)
First-degree
family
mean (N) F
P
MOTHER-
CARING
16.23 (126)
16.23 (126)
16.23 (126)
19.19 ( 37)
19.19 ( 37)
19.19 ( 37)
20.93 (105)
20.93 (105)
20.93 (105)
7.080
2.526
0.971
14.317
.0 0 1 **
.114
.326
.0 0 0 **
MOTHER
OVER-
PROTECTIVE
22.43 (127) 20.10 ( 39) 20.65 (102) 1.580 .208
FATHER-
CARING
18.37 (115)
18.35 (115)
18.35 (115)
21.94 ( 35)
21.94 ( 35)
21.94 ( 35)
23.49 ( 95)
23.49 ( 95)
23.49 ( 95)
8.155
3.992
0.659
16.377
.001**
.048*
.418
.000**
FATHER
OVER-
PROTECTIVE
23.78 (116) 21.97 ( 32) 21.46 ( 93) 1.971 .142
87
Correlational Data. Pearson correlations were computed between variables
describing parental psychological treatment ( d id a p a r e n t h a v e p s y c h o l o g ic a l
TREATMENT? a n d DID EITHER PARENT HAVE A PSYCHIATRIC HOSPITALIZATION?), p a r e n t a l
excessive use of alcohol ( d id e it h e r p a r e n t d r in k a l c o h o l t o e x c e s s ? ), and the
scores on survey instruments (CES-D, Rosenberg-SEI, and PBI) and between
these parental variables and c s a . These are shown in Tables 11 through 13.
As stated previously, because of the number of correlations performed, only
those reaching a level of significance of at least .001 are discussed here.
Those correlations reaching significance at the .001 level were as follows:
1. DID A PARENT HAVE PSYCHOLOGICAL TREATMENT? C o r r e l a t e d
significantly with the f a t h e r -c a r in g scale of the PBI (r = .163;
p = .001). When correlations were performed using only the
comparison (no-CSA) group, the correlation was significant (r
= .215, p < .001). The correlation using data from the group
judged to have experienced CSA was not significant. These
results suggest that for respondents who had a parent who
received psychological treatment, fathers were perceived or
remembered as being more caring.
2 . d id e it h e r p a r e n t d r in k a l c o h o l t o e x c e s s ? correlated
significantly with the m o t h e r -c a r in g subscale of the PBI (r =
88
.288; p < .001). This correlation was also significant when
using data from the group judged to have experienced CSA (r
= .256, p < .001) and the comparison group (r = .269, p <
.001). These correlations suggest that for respondents who had
at least one parent who drank to excess, the mother was
perceived or remembered as more caring.
3. DID EITHER PARENT DRINK ALCOHOL TO EXCESS? C o r r e l a t e d
significantly with the f a t h e r - c a r i n g subscale of the PBI (r =
.191; p < .001). This correlation was also significant when
using data from the group judged to have experienced CSA (r
= .175, p = .001) and the comparison group (r = .156, p = .05).
These correlations suggest that for respondents who had a
parent who drank to excess, the father was also perceived or
remembered as more caring.
4. DID EITHER PARENT DRINK ALCOHOL TO EXCESS? C o r r e l a t e d
significantly with the experience of childhood sexual abuse
( c s a ) (r = .187; p < .000). This suggests that for respondents
who had a parent who drank to excess, there was greater
likelihood of the respondent having experienced childhood
sexual abuse.
89
Earlv Sexual Experiences
Of 491 respondents who answered, 49.9% said they had sexually kissed or
touched someone five or more years older before the age of 17. Of the total
number, 19.7% said that force had been used on at least one occasion, and
29.2% reported intimidation being used.
Four hundred ninety women responded to the question about whether
they had sexual experiences including genital contact before the age of 17.
Of these, 45% said they had such an experience with someone 5 or more
years older. In 21.9% of cases this sexual contact involved physical force,
and in 33.7% it involved intimidation.
Of 489 respondents, 34.3% said they had experienced penetration
(oral, anal or vaginal) before the age of 17 with someone 5 or more years
older. Of the total sample, 19.9% reported physical force used, and 27.4%
reported verbal intimidation.
Respondents were given a list of possible persons with whom they
might have had early sexual experiences and asked to check off those that
applied. For those who were abused, often there was more than one abuser.
Results of this questioning are shown in Table 16.
90
Table 16. Percentages of respondents with CSA histories who reported
sexual experiences with various persons 1
PERSON TYPE OF SEXUAL EXPERIENCE
NON-GENITAL TOUCH GENITAL TOUCH PENETRATION
FRIEND 31.6 26.2 19.1
DATE 39.4 33.7 27.7
STRANGER 14.9 15.6 11.7
FAMILY 2 43.3 41.1 27.3
FATHER 19.1 16.3 1 2 .0
MOTHER 6 .0 5.7 4.3
SIBLING 1 1 .0 13.0 7.8
FAMILY FRIEND 28.0 22.7 14.2
BABYSITTER
7.1 7.8 4.3
PROFESSIONAL 8.5 6 .2 3.9
OTHER 24.1 24.1 15.6
Many respondents reported sexual experiences with more than one person, therefore percentage
totals for each column exceed 1 0 0 %.
Includes extended family
91
In summary, of those who reported sexual abuse as children, 47.0%
said they were abused by a non-family perpetrator, 38.8% by someone in
their first-degree family (father, mother or sibling), and 14.2% by an
extended family member (e.g., uncle or grandfather). The mean number of
abuse events reported by those who were judged to have been sexually
abused was 110.7 (SD = 71.8; median = 124; mode = 183), with 32
participants reporting more than 60 instances of each type of abuse (sexual
kissing, genital touching and penetration). Excluding these 32 people, the
remaining participants who reported a frequency of abuse events reported a
mean number of total abuse events o f 58.2 (SD = 47.86; median = 25.0;
mode = 25.0). Since a participant may have reported a single instance of
abuse in each section if it involved sexual kissing and genital touching and
penetration, the most conservative estimate of the average number of events
for the typical participant would be 58/3 or 19.3 events, with 32 participants
(about 1% of the CSA-group) reporting more than 60 events of each type.
Women who responded positively to questions about sexual
experiences when they were younger than 17 were also asked about whether
or not they had talked to anyone about the experience, or specifically to a
therapist. If they had talked to someone, they were asked if they felt at the
time that they were listened to and believed, and how long it was before they
92
talked to someone. Within the CSA group, 21.3% said they were able to
discuss the experience at the time it happened, and of those who could talk
about it at the time, 51.6% said that they felt they were listened to and
believed when they talked about their experience. Within the CSA group,
80.9% said that they talked about their experience at some time. For the
largest number of these women, it was more than 10 years before they talked
to anyone (41.1%). Thirteen percent said that they talked to someone
immediately, 5.3% in a few days, 3.2% in a few weeks to a month, 2.5% in
two to twelve months, 7.8% in one to five years, and 10.3% in six to ten
years. Of those in the CSA group, 63.1% said that they had talked to a
therapist about the experience, and of these, only 1.2% talked to the therapist
within a year. The majority (78.9%) talked to a therapist more than ten years
later, with 8.4% talking to a therapist one to five years later, and 11.4%
talking to a therapist six to ten years later.
Respondents Evaluations of Childhood Experience
As stated previously, 49.7% of the respondents said that they did not consider
themselves to have been sexually abused as children, 41.0% said they were
sexually abused as children, and 8.7% said that they were "uncertain." Of
the 282 women classified by the definition used in this study as having been
sexually abused in childhood (CSA group), 23% (N=66) said that they did
93
not think they were sexually abused as children. Some (8.5%) were uncertain
about whether they were sexually abused as children, and 67.7% said that
they were sexually abused as children.
Of the total sample, 64.5% reported that they had not experienced
physical abuse, while 28.6% said that they had, and 6.1% said they were not
certain. Of the total sample, only 34.5% said that they had not been either
sexually or physically abused-- with 63.5% stating that they had experienced
abuse of one kind or the other (or both) as children.
Respondents were asked to rate their childhood on a range (7-point
Likert scale) from very unhappy (1) to very happy (7). The mean rating was
3.6 with a median of 3.0 and a mode of 2.0. Responses were distributed as
follows:
1 2 3 4 5 6 7
9.2% 21.0% 19.9% 16.6% 16.6% 11.6% 4.7%
Data Analysis/Research Questions
Comparison of Depressive Symptoms
The mean depressive symptoms score for respondents in the CSA group was
21.83 (N=272), while the mean depressive symptoms score for respondents
94
not in the CSA group was 18.14 (N=199). This difference was significant
(F=12.50; p < .001). This finding suggests that lesbian women who were
judged by the criteria of this study to have been sexually abused in childhood
are, in this sample, more likely to show depressive symptomatology in
adulthood than lesbian women who were not sexually abused as children.
Analysis of Variables as Mediators of the
Relationship Between c s a and
ADULT DEPRESSIVE SYMPTOMS
Results o f the path analysis are shown in Figure 3. Arrows are drawn
between the independent variable ( c s a ) and the intervening variables, between
the independent variable and dependent variable ( a d u l t d e p r e s s iv e s y m p t o m s ),
and between the intervening variables and dependent variable. Path
coefficients (shown next to arrows) indicate strengths of the relationships
between variables.
As shown in Figure 3, s e l f -e s t e e m is highly significant in the
relationship between c s a and a d u l t d e p r e s s iv e s y m p t o m s . In addition, the
m o t h e r -c a r in g variable derived from the PBI is also significant in the
relationship between c s a and a d u l t d e p r e s s iv e s y m p t o m s . These levels of
significance hold whether the full CES-D score is used ( a d u l t d e p r e s s iv e
95
Figure 3. Results from the path analysis examining variables as potential
mediators in the relationship between CSA and a d u l t d e p r e s s iv e
s y m p t o m s . The independent variable ( c s a ) is shown on the left.
Variables in the center column are intervening variables (potential
mediators). The dependent variable ( a d u l t d e p r e s s iv e s y m p t o m s ) is
shown on the left. Values shown are path coefficients. Negative
relationships are drawn with a dashed line. Significance levels are
indicated as follows: a single asterisk (*) p <05; a double
asterisk (**), p < .01; b o ld face type and a double asterisk (**), p
> .001. N = 383
-.004
1
.170** i
J
.236** i
CSA
J
____-^210^_
.149**,
1
-.277** j
SELF-ESTEEM
MOTHER-CARING
/MOTHER
IpVER-PROTECTIVE.
.679**
> .112 *
A -.060 |
y I
ADULT
DEPRESSIVE
SYMPTOMS
FATHER-CARING^ _ _ -.039
/FATHER
loVER-PROTEC
96
s y m p t o m s ) or the shortened version of the CES-D is used ( c e s -d a b b r e v ia t e d ),
with only slight changes in the values of path coefficients (e.g., .112 to .113
for m o t iie r -c a r i n g ; .679 to .644 for s e l f-e s t e e m ; -.039 to -.038 for f a t h e r -
c a r in g ).
The correlation between c s a and a d u l t d e p r e s s iv e s y m p t o m s (shown in
Table 11) is .159 (p < .001). The path coefficient for the relationship
between c s a and a d u l t d e p r e s s iv e s y m p t o m s in the presence of the intervening
variables is reduced to a value of -.004. These data are consistent with a
model in which s e l f -e s t e e m and m o t h e r -c a r in g are acting as mediators.
Indirect Effects
Indirect effects were computed for the significant variables, s e l f -e s t e e m and
m o t h e r -c a r in g (Cohen and Cohen, 1983), These are shown in Table 17. The
indirect effect of c s a on a d u l t d e p r e s s iv e s y m p t o m s through s e l f-e s t e e m is
.131, and of c s a on a d u l t d e p r e s s iv e s y m p t o m s through m o t h e r -c a r in g is .077.
These indirect effects indicate that se l f-e s t e e m and m o t h e r -c a r in g are acting
as partial mediators, with a small direct effect remaining from c s a . This
direct effect from c s a also contributes to the variance in a d u l t d e p r e s s iv e
s y m p t o m s .
97
Table 17. Indirect effects between c s a and partial mediators ( s e l f -e s t e e m and
m o t h e r -c a r in g ) in the relationship between c s a and a d u l t
d e p r e s s iv e s y m p t o m s . Pearson correlations between c s a and the
partial mediator were multiplied with Pearson correlations
between the partial mediator and a d u l t d e p r e s s iv e s y m p t o m s to
obtain the indirect effects. Results further demonstrate that s e l f
e s t e e m and m o t h e r -c a r in g act as partial mediators, with a
remaining direct effect of c s a on a d u l t d e p r e s s iv e s y m p t o m s (r =
.159, p < .001).
Mediator fi Correlation
with c s a
Correlation
with ADULT
DEPRESSIVE
SYMPTOMS
Indirect
Effect
SELF
ESTEEM
.679 .179 .730 .131
MOTHER-
CARING
.1 1 2 .247 .310 .077
98
Additional Analysis
Because s e l f -e s t e e m was seen to account for so much of the variance in
a d u l t d e p r e s s iv e s y m p t o m s (change in R2 of .40 when added to the regression
equation), a subsequent regression was performed using only the variables
derived from the PBI as independent variables. This was done in order to
look at the maximum effect of these variables when considered in the
absence of other potential mediators which might share variance or be
intercorrelated with other variables in the equation. Results are shown in
Figure 4, and demonstrate that in the absence of other potential mediators, the
variables measuring the respondent’s recalled relationship with her mother
( m o t h e r -c a r in g and m o t h e r o v e r -p r o t e c t iv e ) could be partial mediators.
Variables measuring the recalled relationship with the father-figure ( f a t h e r -
c a r in g and f a t h e r o v e r -p r o t e c t iv e ) do not reach significance and are not
acting as mediators.
Summary
Analysis shows significant relationships between c s a and intervening
variables and between s e l f -e s t e e m and m o t h e r -c a r in g and a d u l t d e p r e s s iv e
s y m p t o m s . Together with computed indirect effects, these data are consistent
with a model that proposes that s e l f -e s t e e m is an important partial mediator
99
Figure 4. Path analysis in the absence o f the s e l f - e s t e e m variable. Results
from the path analysis examining PBI variables alone as potential
mediators in the relationship between c s a and a d u l t d e p r e s s i v e
s y m p to m s . The independent variable ( c s a ) is shown on the left.
Variables in the center column are intervening variables (potential
mediators). The dependent variable ( a d u l t d e p r e s s i v e s y m p to m s ) is
shown on the left. Values shown are path coefficients. Negative
relationships are drawn with a dashed line. Significance levels are
indicated as follows: a single asterisk (*) p <05; a double
asterisk (**), p < .01; bold face type and a double asterisk (**), p
>001. N = 385
.060
.235** ,
J
-.208** to
CSA
.r , . J
.147*’ t
1
-.274** «
------------------------------------- j
► <;
MOTHER-CARING 1 .221*
'MOTHER
OVER-PROTECTIVi
""\ - 215**
FATHER-CARING
'FATHER
. OVER-PROTECTI
.014
.022
ADULT
DEPRESSIVE
SYMPTOMS
100
of the relationship between c s a and a d u l t d e p r e s s iv e s y m p t o m s . This remains
true when the shortened version of the CES-D (eliminating questions related
to self-esteem) is used. In addition, data are consistent with the m o t h e r -
c a r in g variable being a partial mediator of the relationship between c s a and
a d u l t d e p r e s s i v e s y m p t o m s , with the participant’s memory of mother as more
caring being associated with higher levels of a d u l t d e p r e s s iv e s y m p t o m s .
Indirect effects computed for the s e l f -e s t e e m and m o t h e r -c a r in g
variables showed each to have indirect effects accounting for part of the
relationship between c s a and a d u l t d e p r e s s iv e s y m p t o m s , while a direct effect
of c sa on a d u l t d e p r e s s iv e s y m p t o m s also remained.
Finally, if PBI variables are examined in the absence of any other
potential mediators (e.g., s e l f -e s t e e m ) only the variables relating to the
mother appear to be partial mediators, with the participant’s recollection of
the mother being more "caring" and more "over-protective" associated with
higher levels o f a d u l t d e p r e s s iv e s y m p t o m s . This is in opposition to previous
findings with regard to the m o t h e r -c a r in g variable. Previous findings have
related subject’s recall of mother as more "caring" with lower levels of
depressive symptoms (Bifulco, Brown and Harris, 1994; Parker, 1983).
Previous research has produced conflicting results regarding the relationship
between the participant’s memory of her mother as "over-protective"
(Bifulco, Brown and Harris, 1994).
101
Analysis of Variables as Moderators of the
Relationship Between c s a and a d u l t d e p r e s s iv e s y m p t o m s
The analyses described in this section were performed to determine whether
the variables being studied here ( s e l f -e s t e e m , PBI subscales, and d id y o u t a l k
t o a t h e r a p is t ?) act as moderators in the relationship between c s a and a d u l t
d e p r e s s iv e s y m p t o m s . First, all variables were centered by converting to z-
scores. Then, a regression equation was constructed for each potential
moderator. The potential moderator variable was entered into the equation
first, followed by c s a and then an interaction variable computed by
multiplying the c s a value for each case by the value of the potential
moderator. If the interaction variable is acting as a moderator in relationship
between c s a and a d u l t d e p r e s s iv e s y m p t o m s , it is expected to account for a
significant amount of unique variance in a d u l t d e p r e s s iv e s y m p t o m s . Results
of this analysis are shown in Tables 18 through 20. In no case did the
interaction term account for a significant amount of variance in a d u l t
d e p r e s s iv e s y m p t o m s , indicating that the independent variables in question do
not act as moderators in the relationship between a d u l t d e p r e s s iv e s y m p t o m s
and c s a .
102
Table 18. Moderator analysis using the s e l f -e s t e e m variable, s e l f -e s t e e m is
shown to account for a significant (53.1%) amount o f variance in
a d u l t d e p r e s s iv e s y m p t o m s . When c s a is subsequently added to the
equation, no significant variance is accounted for. In addition, the
interaction between c s a and s e l f -e s t e e m does not account for
significant variance in a d u l t d e p r e s s iv e s y m p t o m s indicating that
the s e l f -e s t e e m variable is not acting as a moderator.
N
change
R2
p change
R2 B SE B 1 3
SELF-ESTEEM 467 .531 .0 0 0 ** .731 .033 .721
CSA .0 0 1 .266 .039 .032 .039
Interaction .0 0 1 .367 .030 .033 .029
103
Table 19. Moderator analysis using the PBI variables. All four PBI
variabnles are shown to account for a significant variance in a d u l t
d e p r e s s iv e s y m p t o m s. When c s a is subsequently added to the
equation, it accounts for significant variance in a d u l t d e p r e s s iv e
s y m p t o m s in three cases ( m o t h e r o v e r -p r o t e c t iv e , f a t h e r -c a r in g
and f a t h e r o v e r -p r o t e c t iv e ) indicating a significant direct effect
of the c s a variable. In all four cases the interaction between c s a
and the PBI variables does not account for significant variance in
a d u l t d e p r e s s iv e s y m p t o m s indicating that these variables are not
moderating the relationship between c s a and a d u l t d e p r e s s iv e
s y m p t o m s .
N
change
R2
p change
R2 B SE B B
MOTHER-CARING 447 .096 .0 0 0 ** .289 .046 .288
CSA .006 .077 .091 .046 .091
Interaction .006 .078 .082 .047 .082
MOTHER OV ER-
PROTECTIVE 450 .096 .0 0 0 ** -.294 .046 -.290
CSA .013 .0 1 1 * .118 .046 .118
Interaction .0 0 0 .637 - .0 2 2 .046 - .0 2 1
FATHER-CARING 415 .035 .0 0 1 ** .161 .049 .161
CSA .018 .006** .134 .049 .134
Interaction .0 0 0 .740 -.016 .049 -.016
FATHER OVER-
PROTECTIVE 412 .045 .0 0 0 ** -.184 .051 -.181
CSA .014 .014* .1 2 2 .051 .1 2 1
Interaction .0 0 0 .907 .006 .052 .006
104
Table 20. Moderator analysis using the variable, d id y o u t a l k t o a
THERAPIST?. DID YOU TALK TO A THERAPIST? is s h o w n tO aCCOUnt f o r a
significant ( 6 . 1 % ) amount of variance in a d u l t d e p r e s s iv e
s y m p t o m s . When c s a is subsequently added to the equation, no
significant variance is accounted for. In addition, the interaction
between c s a and d id y o u t a l k t o a t h e r a p is t ? does not account for
significant variance in a d u l t d e p r e s s iv e s y m p t o m s indicating that
the d id y o u t a l k t o a t h e r a p is t ? variable is not acting as a
moderator.
N
change
R2
p change
R2 B SE B a
DID YOU TALK
TO A THERAPIST? 428 .061 .000** .2 0 2 .054 .198
CSA .006 .094 .106 .056 .1 0 2
Interaction .0 0 2 .301 .058 .056 .051
105
Summary: Research Questions
Are lesbian women who report having been sexually abused in childhood
more likely to show depressive symptomatology in adulthood than lesbian
women who did not report that they were sexually abused as children?
Correlational data demonstrate that, in this sample, lesbian woman who were
judged by the survey criteria to have been sexually abused as children
experience higher levels of a d u l t d e p r e s s i v e s y m p to m s than lesbian women
judged not to have experienced CSA.
Does s e l f - e s t e e m mediate or moderate the relationship between CSA and
adult depressive symptoms?
Path analysis is consistent with a model in which s e l f -e s t e e m is an important
partial mediator of the relationship between c s a and a d u l t d e p r e s s iv e
s y m p t o m s , s e l f -e s t e e m was not shown to moderate the relationship between
c s a and a d u l t d e p r e s s iv e s y m p t o m s .
106
Does the respondent’ s retrospective report of parental bonding mediate or
moderate the relationship between c s a and depressive symptoms in
adulthood?
Path analysis is consistent with a model in which the variable measuring the
respondent’s recollection of her mother as "caring" ( m o t h e r -c a r in g ) is a
partial mediator of the relationship between c s a and a d u l t d e p r e s s iv e
s y m p t o m s . The results suggest that a memory of the mother as more caring is
associated with higher levels of depressive symptoms. The m o t h e r -c a r in g
variable does not moderate the relationship between c s a and a d u l t d e p r e s s iv e
s y m p t o m s .
When PBI variables alone are entered into the path analysis (leaving
out s e l f -e s t e e m ) the m o t h e r o v e r -p r o t e c t iv e variable also appears as a partial
mediator of the relationship between c s a and a d u l t d e p r e s s iv e s y m p t o m s . In
this case a memory of the mother as more "over-protective" is associated
with lower levels of adult depressive symptoms. The m o t h e r o v e r -p r o t e c t iv e
variable does not appear to moderate the relationship between c s a and a d u l t
DEPRESSIVE SYMPTOMS.
Variables having to do with the respondent’s memory of the father-
figure ( f a t h e r -c a r in g and f a t h e r o v e r -p r o t e c t iv e ) did not reach significance
in either the mediator or moderator analysis, and do not act as mediators or
107
moderators of the relationship between c s a and a d u l t d e p r e s s iv e s y m p t o m s .
In all significant mediational results, variables ( s e l f -e s t e e m and
m o t h e r -c a r in g ) act as partial mediators, with a direct effect remaining
between c s a and a d u l t d e p r e s s iv e s y m p t o m s .
Does participation in psychotherapy moderate the effects of c s a ?
Data did not show participation in psychotherapy to moderate the relationship
between c s a and a d u l t d e p r e s s iv e s y m p t o m s . However, the data collection was
not sufficient to fully address this question.
108
DISCUSSION
In this sample of lesbian women, women who reported events in their
childhood interpreted to be sexually abusive were shown to experience more
adult symptoms of depression, or experience adult depressive symptoms more
strongly than do women who did not report such events. This result is
consistent with the literature on long-term effects of CSA on women in
general, which often cites depression as a long-term correlate of CSA (e.g.,
Bifulco, Brown and Adler, 1991; Briere, 1992a&b; Gamefski, van Egmond
and Straatman, 1990; Gelinas, 1983; Herman, 1992; Koverola, Pound, Heger
and Lytle, 1993). Although no causal relationship can be definitively
determined from these data (Cliff, 1983) the data are consistent with a causal
model in which self-esteem and the respondent’s perception of her mother as
more caring act as mediators in the relationship between the experience of
childhood sexual abuse and adult depressive symptoms.
Some respondents in this study reported that they did or did not
experience CSA in opposition to the judgement based on the criteria used in
109
the study to determine whether respondents experienced CSA or not. For
example, 8 participants stated that they were sexually abused as children,
while they did not meet the criteria for CSA used in the study. Analysis was
done to determine whether there were any significant differences on any of
the "outcome variables" ( s e l f -e s t e e m , m o t h e r -c a r in g , m o t h e r o v e r -p r o t e c t iv e ,
FATHER-CARING, FATHER OVER-PROTECTIVE, O r ADULT DEPRESSIVE SYMPTOMS) b e t w e e n
respondents who agreed and who disagreed with the classification based on
the study criteria. For those groups of respondents who agreed with the
study determination about their abuse status, significant differences were
found, for all variables tested, between the CSA and non-CSA groups. The
salient finding in comparing groups of women who disagreed with the
judgement based on study criteria that they had experienced CSA is that they
saw their parents as significantly less caring than those who agreed with the
study determination that they had experienced CSA. This leads to some
speculation (not addressed in this study) about the value of seeing these
experiences as abusive, and the possibly detrimental effects on relationships
with parents of denying that the early sexual experiences were abusive.
Path analysis performed in this study is consistent with a model
incorporating self-esteem as an important mediator of the relationship
between c s a and a d u l t d e p r e s s iv e s y m p t o m s in lesbian women. Self-esteem is
sometimes seen to be part of depression, and efforts have been made here and
elsewhere to determine how closely related the measures used to asses s e l f -
e s t e e m and a d u l t d e p r e s s iv e s y m p t o m s are (Orme, Reis and Herz, 1986).
However, research has not shown the two constructs to be the same, and it is
possible that the sexual abuse may be a causal agent in the development of
low self-esteem, and that the low self-esteem in turn may play a part in the
development of depressive symptomatology as a whole, and not just the parts
of depressive symptomatology that have to do directly with negative self-
image. Briere (1992a) has suggested that cognitive sequelae of CSA may
contribute to or mediate negative symptoms in adults, and that the negative
self-perceptions associated with CSA may be due to coexisting psychological
abuse that is inherent in all types of abuse. These negative self-perceptions
may be reflected in responses to the self-esteem measure used in this study.
Correlational analysis of the four PBI subscales and the a d u l t
d e p r e s s iv e s y m p t o m s variable showed a positive relationship between this
variable and the two "caring" subscales, and a negative relationship between
a d u l t d e p r e s s iv e s y m p t o m s and the two "over-protection" subscales. This
result raises a number of questions, among them: (1) Whether our view of
caring and supervised independence as elementw of good parenting leading to
strong attachment and well adjusted offspring should be called into question,
or (2) Whether the scales are accurately named and are actually measuring
caring and over-protection. For example, is "over-protection" as measured
here really a constraining and overly intrusive phenomenon, or does this scale
actually measure parental "protection" which was at the time defined by the
child as (or is currently remembered by the grown child as) too intrusive and
allowing too little freedom. There are also questions raised about the type
and extent of distortion which may be present because of the retrospective
nature of these scales. These questions will be addressed more fully below.
Analysis of the relationships between c s a , a d u l t d e p r e s s iv e s y m p t o m s
and the four subscales of the PBI shows of the four variables derived from
the PBI, only the m o t h e r -c a r in g variable acts as a partial mediator of the
relationship between c s a and a d u l t d e p r e s s iv e s y m p t o m s . When s e l f -e s t e e m is
removed from the path analysis (in order to see the maximal effects of the
PBI variables), only the two variables relating to the mother ( m o t h e r -c a r in g
and m o t h e r o v e r -p r o t e c t iv e ) act as partial mediators. The signs of the partial
regression coefficients demonstrate that perception of the mothers to have
been more caring is associated with higher levels of a d u l t d e p r e s s iv e
s y m p t o m s , and perception of the mother as more "over-protective" are here
associated with lower levels of a d u l t d e p r e s s iv e s y m p t o m s . The association of
m o t h e r -c a r in g and higher a d u l t d e p r e s s iv e s y m p t o m s is in opposition to
existing literature (Bifulco, Brown and Harris, 1994; Kendall-Tackett,
Williams and Finkelhor, 1993; Parker, 1983; Parker, Tupling and Brown,
1979). The association of the m o t h e r o v e r -p r o t e c t iv e variable and lower
levels of a d u l t d e p r e s s iv e s y m p t o m s has not been consistently reported
(Bifulco, Brown and Harris, 1994).
It appears from these data that the behavior of the mother is more apt
to influence the relationship between c s a and a d u l t d e p r e s s iv e s y m p t o m s than
that of the father. This finding is consistent with thek work of Kendall-
Tackett, Williams and Finkelhor (1993) who found the mother to be more
influential. However, findings with these data show that a perception of the
mother as having been more caring is somehow associated with higher levels
of adult depressive symptomatology. This is in opposition to the findings of
Kendal-Tackett, Williams and Finkelhor (1993), who found "support" from
the mother to be protective. The influence of the mother may be related to
the fact that the mother (as reported in these data) is much less often the
perpetrator of abuse. Perhaps, it the fact of being Mother is not as salient as
is the fact of being the non-abusing parent or the parent least like the abuser.
Additional data would be necessary to test this hypothesis, and further
research needs to be done to elucidate the relationship between the perception
of the mother as caring and higher levels of adult depressive symptoms.
These findings are important because they comprise one of the first
examinations of the relationship between parental bonding and negative
symptoms in adult survivors of CSA. There is currently a significant
literature in which theoretical propositions suggest a variety of possible
113
relationships between CSA, the quality of family life and adult negative
symptoms such as depression. A few writers in this field present data and
theory in a manner suggesting that the family influence is so important that
the CSA itself becomes unimportant or non-contributory to the adult
symptomatology. It is crucial, when thinking about these relationships, to
bear in mind that these data are consistent with a model in which c s a also has
a direct effect on adult symptoms of depression. Certainly the relationships
between these variables are not simple, and a vast number of models could be
proposed and confirmed. Even then, the confirmation does not prove
causality (Cliff, 1983).
In order to sort out some of the complexities the respondents who
were judged to have been sexually abused as children were categorized
according to the closest perpetrator of that abuse (first degree family,
extended family, or non-family). ANOVA demonstrated that there were
significant differences in the both the m o t h e r -c a r in g and f a t h e r -c a r in g scores
for women who were abused by a first degree family member and women
who were abused by someone outside the family. Those reporting abuse by
someone outside the family saw their parents as significantly more caring
than those who reported abuse by a first degree family member. This further
suggests that future research should examine more closely the relationships
between the variables studied here in light of the particular aspects of the
abuse. The specific relationship of the perpetrator to the child would be an
important area for further investigation. For example, does the fact that a
perpetrator was a parent change the perception of that parent as caring~or are
less caring parents more likely to perpetrate abuse. One could also inquire
about whether variations in established relationships with parents are
associated with more adult depressive symptoms in the presence of CSA,
than they would be in the absence of CSA. These and many more questions
must be answered before the complexity of these relationships can be
elucidated.
The final research question in this study addresses the effect of
psychotherapy on a d u l t d e p r e s s iv e s y m p t o m s . It was found that respondents
to this survey who talked to a therapist at sometime reported experiencing
more depressive symptoms or more severe symptoms than those who did not.
This finding could suggest a number of interpretations, but it seems
reasonable to conclude that those who were experiencing more depressive
symptoms may have sought out more psychotherapeutic help. Further study
would need to be done to determine whether this hypothesis is valid, and
whether the therapy is helpful. These are questions which cannot be
answered with these data.
One of the difficulties with this study, in general, is the retrospective
nature of all measures. For example, the CES-D asks the participant to recall
115
how she felt in the past week; the Early Sexual Experiences questionnaire
asks for a description of highly charged events sometimes long past; and the
PBI asks her to go back to her earliest memories of her parents. These
evaluations of early experiences with parents include answering questions
about how warm, friendly, helpful, or constraining (for example) the
respondent’s parents were. It seems likely that respondent’s evaluations
could be greatly influenced by current and past attributions and affective
states. In essence, this study has not measured the attachment between the
respondent and her parents, but the respondent’s current evaluation and
memory of that attachment.
One might ask whether retrospection may be more influential or
influential in different ways with more and less emotionally laden and
emotionally sensitive material. All of the questions in this study have
emotional components, but it may be qualitatively different to be asked "were
you sexually abused?" or "did anyone touch you in a sexual/genital way" than
it is to be asked to rate whether you are a "good person" or to rate how warm
or helpful or restrictive a parent was.
There is evidence in these data that the women who reported abuse by
first degree relatives perceived their parents to be less caring than those who
reported abuse by non-family members. However, the presence of abuse
trauma may also, itself, influence recollections by creating a type of
attachment or bonding, although not necessarily a healthy one. This reaction
has been suggested to occur in some abusive relationships between adults
(Grand and Alpert, 1993) and has been labeled "traumatic bonding" by
DeYoung and Lowry (1992). Such bonding was demonstrated in a study of
75 women in abusive relationships by Dutton and Painter (1993). Other
authors have suggested that the abused child may be viewed as a hostage and,
further, that literature on the behavior of inmates in concentration camps and
the "Stockholm Syndrome" may apply to children in abusive situations
(Goddard and Stanley, 1994). Such attachment might affect responses to an
instrument like the PBI. A possible result might be that abused respondents
would appear to be in relationships with their parents characterized by high
levels of bonding or attachment, when what is being measured is the holding
power of trauma and threat— not the parental attachment described by
researchers such as Bowlby and Ainsworth. It is important here to do more
work so we know we are measuring the appropriate phenomena.
A similar criticism can be raised to the use of the CES-D which asks
whether the respondent has experienced symptoms in the past week. It is
possible that the respondent’s affect for the week has been influenced by
some current life difficulty such as the death of parent or the loss of a job.
Critics suggest that currently felt negative emotions can cause an
inappropriately high score on the instrument, which is then interpreted as a
general state of depression rather than a specific situational case. In fact, a
number of respondents wrote in the margins of survey forms that they had
just lost a job or a lover had just left and they were concerned that their
answers would paint an inappropriately bleak picture. One of the
respondents, however, wrote in the margin that she was worried that the
scores on depression questions would be artificially low (few symptoms)
because numbers of women were filling out the form during "Women’s
Week" celebrations in Provincetown, MA. An estimated 28% (N=140) of the
respondents to this survey were in unusually positive and supportive
situations such as "Women’s Week" or the "Coming Out Day Celebration" in
Pasadena, CA when they filled out the survey forms. It appeared, from these
comments in the margins and given a knowledge of the number of women
filling out the form at celebratory (and presumably happy) events, that people
are much more willing to say "it isn’t really that bad, I’m having a bad
week," than to say "it isn’t usually this good." There seems to be no reason
to believe that more people are having unusually bad weeks than are having
unusually good weeks. In conclusion, it is unknown whether any non
surveyed factors were influencing the CES-D score, but it is reasonable to
believe that there was as much or more influence in the non-depressive
direction than in the depressive direction.
The retrospective nature of reporting early sexual experiences and
abusive sexual experiences is also problematic. In relation to CSA all of the
difficulties described above apply (such as the possibilities of current affect
or the affect elicited at the time of the event influencing memory). In
addition, there may be a further difficulty posed by the possibility that
participants might not remember, and therefore not report, incidents of CSA.
These people would then be erroneously placed in the comparison group and
might act to reduce the effect size observed in analysis. Warnings about such
possible "noise" in the comparison group— caused by people who were abused
reporting no abuse— have been issued by Schetky (1990), Briere (1992b), and
Herman and Schatzow (1987). This phenomenon has been investigated by
Briere and Conte (1993) and Williams (1994a) with data demonstrating that
the possibility of amnesia about prior traumatic abuse is a valid concern.
Recently this topic has become quite controversial with heated debate
regarding the possibility or impossibility of not remembering a traumatic
abuse event (e.g. Terr, 1988, 1991; Wakefield and Underwager, 1992).
Further discussion on the topic can be found in papers by Loftus, Garry and
Feldmen (1994) and Williams (1994a&b), among others.
Other researchers have suggested that non-sexual trauma such as
physical abuse in childhood also can contribute to depressive symptoms in
adulthood (e.g., Briere, 1992a). Recently data have been presented which
119
supports this suggestion and further shows that multiple traumas in childhood
may contribute more strongly than a single trauma to depressive symptoms in
adulthood (Fox and Gilbert, 1994; Gidycz et al, 1993; Wind and Silvern,
1992). If this is the case (as it most probably is) one should look
simultaneously at non-sexual childhood trauma in order to obtain a fuller
picture of the etiology of adult depressive symptoms.
Other difficulties with this research were discussed in the introduction
and include such issues as subject selection. These merit further thought and
research. For example, doing a similar survey which concentrates on rural
populations of lesbian women, or imbedding questions in a larger survey of
women so participants can be selected based on reports of sexual behavior
rather than self-identification as lesbian. Such further research would be an
important contribution to this literature.
Additionally, there is the ever-present difficulty surrounding the
definition of sexual abuse. This study definition places 16.2% of the women
who say they were not abused in the abused category. There may be women,
like the one mentioned who stated that she had been married at 16 to
someone 11 years older, who do not see their sexual experience as anything
but normal within the parameters of their life or culture. It is possible that
these cases represent real abuse, or that they are inappropriately placed in the
group (or some of both). It is important to develop a standard definition o f
CSA, such as the one used in this research, which can be used in all research
studies, so the studies can be better compared. On the other hand, it is
important to continue thinking about the effects and implications of the
definition we have chosen.
Finally, this study looks at depressive symptoms, self-esteem and
parental bonding specifically in a sample of lesbian women. It will be
important to do similar studies in other groups of women (e.g. heterosexual
women) to see if the conclusions here apply more generally or are specific to
this group of women.
Although a mediational relationship is shown here between maternal
attributes as measured by the PBI and the presence of adult depressive
symptoms in these women who experienced CSA, it must be remembered
that there is also a direct effect of the abuse demonstrated by these data.
Certainly ways to further attachment between the parents and the child may
have a powerful mediational effect and should be considered. At the same
time, efforts to prevent the abuse altogether should be pursued. These data
also suggest that if CSA has occurred, efforts to increase self-esteem in the
post-trauma child might also prove quite helpful in terms of preventing or
reducing long-term depressive symptoms. In addition, further research is
required to further elucidate the relationship between these variables and the
mechanisms involved.
121
DISCUSSION
In this sample of lesbian women, women who reported events in their
childhood interpreted to be sexually abusive were shown to experience more
adult symptoms of depression, or experience adult depressive symptoms more
strongly than do women who did not report such events. This result is
consistent with the literature on long-term effects o f CSA on women in
general, which often cites depression as a long-term correlate o f CSA (e.g.,
Bifulco, Brown and Adler, 1991; Briere, 1992a&b; Garnefski, van Egmond
and Straatman, 1990; Gelinas, 1983; Herman, 1992; Koverola, Pound, Heger
and Lytle, 1993). Although no causal relationship can be definitively
determined from these data (Cliff, 1983) the data are consistent with a causal
model in which self-esteem and the respondent’s perception of her mother as
more caring act as mediators in the relationship between the experience of
childhood sexual abuse and adult depressive symptoms.
Some respondents in this study reported that they did or did not
experience CSA in opposition to the judgement based on the criteria used in
109
the study to determine whether respondents experienced CSA or not. For
example, 8 participants stated that they were sexually abused as children,
while they did not meet the criteria for CSA used in the study. Analysis was
done to determine whether there were any significant differences on any of
the "outcome variables" ( s e l f -e s t e e m , m o t h e r -c a r in g , m o t h e r o v e r -p r o t e c t iv e ,
FATHER-CARING, FATHER OVER-PROTECTIVE, O r ADULT DEPRESSIVE SYMPTOMS) b e t w e e n
respondents who agreed and who disagreed with the classification based on
the study criteria. For those groups of respondents who agreed with the
study determination about their abuse status, significant differences were
found, for all variables tested, between the CSA and non-CSA groups. The
salient finding in comparing groups of women who disagreed with the
judgement based on study criteria that they had experienced CSA is that they
saw their parents as significantly less caring than those who agreed with the
study determination that they had experienced CSA. This leads to some
speculation (not addressed in this study) about the value of seeing these
experiences as abusive, and the possibly detrimental effects on relationships
with parents of denying that the early sexual experiences were abusive.
Path analysis performed in this study is consistent with a model
incorporating self-esteem as an important mediator of the relationship
between c s a and a d u l t d e p r e s s iv e s y m p t o m s in lesbian women. Self-esteem is
sometimes seen to be part of depression, and efforts have been made here and
110
elsewhere to determine how closely related the measures used to asses s e l f
e s t e e m and a d u l t d e p r e s s iv e s y m p t o m s are (Orme, Reis and Herz, 1986).
However, research has not shown the two constructs to be the same, and it is
possible that the sexual abuse may be a causal agent in the development of
low self-esteem, and that the low self-esteem in turn may play a part in the
development of depressive symptomatology as a whole, and not just the parts
o f depressive symptomatology that have to do directly with negative self-
image. Briere (1992a) has suggested that cognitive sequelae of CSA may
contribute to or mediate negative symptoms in adults, and that the negative
self-perceptions associated with CSA may be due to coexisting psychological
abuse that is inherent in all types of abuse. These negative self-perceptions
may be reflected in responses to the self-esteem measure used in this study.
Correlational analysis of the four PBI subscales and the a d u l t
d e p r e s s iv e s y m p t o m s variable showed a positive relationship between this
variable and the two "caring" subscales, and a negative relationship between
a d u l t d e p r e s s iv e s y m p t o m s and the two "over-protection" subscales. This
result raises a number of questions, among them: (1) Whether our view of
caring and supervised independence as elementw of good parenting leading to
strong attachment and well adjusted offspring should be called into question,
or (2) Whether the scales are accurately named and are actually measuring
caring and over-protection. For example, is "over-protection" as measured
here really a constraining and overly intrusive phenomenon, or does this scale
actually measure parental "protection" which was at the time defined by the
child as (or is currently remembered by the grown child as) too intrusive and
allowing too little freedom. There are also questions raised about the type
and extent of distortion which may be present because of the retrospective
nature of these scales. These questions will be addressed more fully below.
Analysis of the relationships between c s a , a d u l t d e p r e s s iv e s y m p t o m s
and the four subscales of the PBI shows of the four variables derived from
the PBI, only the m o t h e r -c a r in g variable acts as a partial mediator of the
relationship between c s a and a d u l t d e p r e s s iv e s y m p t o m s . When s e l f -e s t e e m is
removed from the path analysis (in order to see the maximal effects of the
PBI variables), only the two variables relating to the mother ( m o t h e r -c a r in g
and m o t h e r o v e r -p r o t e c t iv e ) act as partial mediators. The signs of the partial
regression coefficients demonstrate that perception of the mothers to have
been more caring is associated with higher levels of a d u l t d e p r e s s iv e
s y m p t o m s , and perception of the mother as more "over-protective" are here
associated with lower levels of a d u l t d e p r e s s iv e s y m p t o m s . The association of
m o t h e r -c a r in g and higher a d u l t d e p r e s s iv e s y m p t o m s is in opposition to
existing literature (Bifulco, Brown and Harris, 1994; Kendall-Tackett,
Williams and Finkelhor, 1993; Parker, 1983; Parker, Tupling and Brown,
1979). The association of the m o t h e r o v e r -p r o t e c t iv e variable and lower
levels of a d u l t d e p r e s s iv e s y m p t o m s has not been consistently reported
(Bifulco, Brown and Harris, 1994).
It appears from these data that the behavior of the mother is more apt
to influence the relationship between c s a and a d u l t d e p r e s s iv e s y m p t o m s than
that of the father. This finding is consistent with thek work of Kendall-
Tackett, Williams and Finkelhor (1993) who found the mother to be more
influential. However, findings with these data show that a perception of the
mother as having been more caring is somehow associated with higher levels
of adult depressive symptomatology. This is in opposition to the findings of
Kendal-Tackett, Williams and Finkelhor (1993), who found "support" from
the mother to be protective. The influence of the mother may be related to
the fact that the mother (as reported in these data) is much less often the
perpetrator of abuse. Perhaps, it the fact of being Mother is not as salient as
is the fact of being the non-abusing parent or the parent least like the abuser.
Additional data would be necessary to test this hypothesis, and further
research needs to be done to elucidate the relationship between the perception
of the mother as caring and higher levels of adult depressive symptoms.
These findings are important because they comprise one of the first
examinations o f the relationship between parental bonding and negative
symptoms in adult survivors of CSA. There is currently a significant
literature in which theoretical propositions suggest a variety of possible
113
relationships between CSA, the quality of family life and adult negative
symptoms such as depression. A few writers in this field present data and
theory in a manner suggesting that the family influence is so important that
the CSA itself becomes unimportant or non-contributory to the adult
symptomatology. It is crucial, when thinking about these relationships, to
bear in mind that these data are consistent with a model in which c s a also has
a direct effect on adult symptoms of depression. Certainly the relationships
between these variables are not simple, and a vast number of models could be
proposed and confirmed. Even then, the confirmation does not prove
causality (Cliff, 1983).
In order to sort out some of the complexities the respondents who
were judged to have been sexually abused as children were categorized
according to the closest perpetrator of that abuse (first degree family,
extended family, or non-family). ANOVA demonstrated that there were
significant differences in the both the m o t h e r -c a r in g and f a t h e r -c a r in g scores
for women who were abused by a first degree family member and women
who were abused by someone outside the family. Those reporting abuse by
someone outside the family saw their parents as significantly more caring
than those who reported abuse by a first degree family member. This further
suggests that future research should examine more closely the relationships
between the variables studied here in light of the particular aspects of the
abuse. The specific relationship of the perpetrator to the child would be an
important area for further investigation. For example, does the fact that a
perpetrator was a parent change the perception of that parent as caring— or are
less caring parents more likely to perpetrate abuse. One could also inquire
about whether variations in established relationships with parents are
associated with more adult depressive symptoms in the presence of CSA,
than they would be in the absence of CSA. These and many more questions
must be answered before the complexity o f these relationships can be
elucidated.
The final research question in this study addresses the effect of
psychotherapy on a d u l t d e p r e s s iv e s y m p t o m s . It was found that respondents
to this survey who talked to a therapist at sometime reported experiencing
more depressive symptoms or more severe symptoms than those who did not.
This finding could suggest a number of interpretations, but it seems
reasonable to conclude that those who were experiencing more depressive
symptoms may have sought out more psychotherapeutic help. Further study
would need to be done to determine whether this hypothesis is valid, and
whether the therapy is helpful. These are questions which cannot be
answered with these data.
One o f the difficulties with this study, in general, is the retrospective
nature of all measures. For example, the CES-D asks the participant to recall
115
how she felt in the past week; the Early Sexual Experiences questionnaire
asks for a description of highly charged events sometimes long past; and the
PBI asks her to go back to her earliest memories of her parents. These
evaluations of early experiences with parents include answering questions
about how warm, friendly, helpful, or constraining (for example) the
respondent’s parents were. It seems likely that respondent’s evaluations
could be greatly influenced by current and past attributions and affective
states. In essence, this study has not measured the attachment between the
respondent and her parents, but the respondent’s current evaluation and
memory of that attachment.
One might ask whether retrospection may be more influential or
influential in different ways with more and less emotionally laden and
emotionally sensitive material. All of the questions in this study have
emotional components, but it may be qualitatively different to be asked "were
you sexually abused?" or "did anyone touch you in a sexual/genital way" than
it is to be asked to rate whether you are a "good person" or to rate how warm
or helpful or restrictive a parent was.
There is evidence in these data that the women who reported abuse by
first degree relatives perceived their parents to be less caring than those who
reported abuse by non-family members. However, the presence of abuse
trauma may also, itself, influence recollections by creating a type of
attachment or bonding, although not necessarily a healthy one. This reaction
has been suggested to occur in some abusive relationships between adults
(Grand and Alpert, 1993) and has been labeled "traumatic bonding" by
DeYoung and Lowry (1992). Such bonding was demonstrated in a study of
75 women in abusive relationships by Dutton and Painter (1993). Other
authors have suggested that the abused child may be viewed as a hostage and,
further, that literature on the behavior of inmates in concentration camps and
the "Stockholm Syndrome" may apply to children in abusive situations
(Goddard and Stanley, 1994). Such attachment might affect responses to an
instrument like the PBI. A possible result might be that abused respondents
would appear to be in relationships with their parents characterized by high
levels of bonding or attachment, when what is being measured is the holding
power of trauma and threat— not the parental attachment described by
researchers such as Bowlby and Ainsworth. It is important here to do more
work so we know we are measuring the appropriate phenomena.
A similar criticism can be raised to the use of the CES-D which asks
whether the respondent has experienced symptoms in the past week. It is
possible that the respondent’s affect for the week has been influenced by
some current life difficulty such as the death of parent or the loss of a job.
Critics suggest that currently felt negative emotions can cause an
inappropriately high score on the instrument, which is then interpreted as a
general state of depression rather than a specific situational case. In fact, a
number of respondents wrote in the margins of survey forms that they had
just lost a job or a lover had just left and they were concerned that their
answers would paint an inappropriately bleak picture. One of the
respondents, however, wrote in the margin that she was worried that the
scores on depression questions would be artificially low (few symptoms)
because numbers of women were filling out the form during "Women’s
Week" celebrations in Provincetown, MA. An estimated 28% (N=140) of the
respondents to this survey were in unusually positive and supportive
situations such as "Women’s Week" or the "Coming Out Day Celebration" in
Pasadena, CA when they filled out the survey forms. It appeared, from these
comments in the margins and given a knowledge of the number of women
filling out the form at celebratory (and presumably happy) events, that people
are much more willing to say "it isn’t really that bad, I’m having a bad
week," than to say "it isn’t usually this good." There seems to be no reason
to believe that more people are having unusually bad weeks than are having
unusually good weeks. In conclusion, it is unknown whether any non
surveyed factors were influencing the CES-D score, but it is reasonable to
believe that there was as much or more influence in the non-depressive
direction than in the depressive direction.
The retrospective nature of reporting early sexual experiences and
abusive sexual experiences is also problematic. In relation to CSA all of the
difficulties described above apply (such as the possibilities of current affect
or the affect elicited at the time of the event influencing memory). In
addition, there may be a further difficulty posed by the possibility that
participants might not remember, and therefore not report, incidents of CSA.
These people would then be erroneously placed in the comparison group and
might act to reduce the effect size observed in analysis. Warnings about such
possible "noise" in the comparison group— caused by people who were abused
reporting no abuse-have been issued by Schetky (1990), Briere (1992b), and
Herman and Schatzow (1987). This phenomenon has been investigated by
Briere and Conte (1993) and Williams (1994a) with data demonstrating that
the possibility of amnesia about prior traumatic abuse is a valid concern.
Recently this topic has become quite controversial with heated debate
regarding the possibility or impossibility of not remembering a traumatic
abuse event (e.g. Terr, 1988, 1991; Wakefield and Underwager, 1992).
Further discussion on the topic can be found in papers by Loftus, Garry and
Feldmen (1994) and Williams (1994a&b), among others.
Other researchers have suggested that non-sexual trauma such as
physical abuse in childhood also can contribute to depressive symptoms in
adulthood (e.g., Briere, 1992a). Recently data have been presented which
119
supports this suggestion and further shows that multiple traumas in childhood
may contribute more strongly than a single trauma to depressive symptoms in
adulthood (Fox and Gilbert, 1994; Gidycz et al, 1993; Wind and Silvern,
1992). If this is the case (as it most probably is) one should look
simultaneously at non-sexual childhood trauma in order to obtain a fuller
picture of the etiology of adult depressive symptoms.
Other difficulties with this research were discussed in the introduction
and include such issues as subject selection. These merit further thought and
research. For example, doing a similar survey which concentrates on rural
populations of lesbian women, or imbedding questions in a larger survey of
women so participants can be selected based on reports of sexual behavior
rather than self-identification as lesbian. Such further research would be an
important contribution to this literature.
Additionally, there is the ever-present difficulty surrounding the
definition of sexual abuse. This study definition places 16.2% of the women
who say they were not abused in the abused category. There may be women,
like the one mentioned who stated that she had been married at 16 to
someone 11 years older, who do not see their sexual experience as anything
but normal within the parameters of their life or culture. It is possible that
these cases represent real abuse, or that they are inappropriately placed in the
group (or some of both). It is important to develop a standard definition of
CSA, such as the one used in this research, which can be used in all research
studies, so the studies can be better compared. On the other hand, it is
important to continue thinking about the effects and implications of the
definition we have chosen.
Finally, this study looks at depressive symptoms, self-esteem and
parental bonding specifically in a sample of lesbian women. It will be
important to do similar studies in other groups of women (e.g. heterosexual
women) to see if the conclusions here apply more generally or are specific to
this group o f women.
Although a mediational relationship is shown here between maternal
attributes as measured by the PBI and the presence o f adult depressive
symptoms in these women who experienced CSA, it must be remembered
that there is also a direct effect of the abuse demonstrated by these data.
Certainly ways to further attachment between the parents and the child may
have a powerful mediational effect and should be considered. At the same
time, efforts to prevent the abuse altogether should be pursued. These data
also suggest that if CSA has occurred, efforts to increase self-esteem in the
post-trauma child might also prove quite helpful in terms of preventing or
reducing long-term depressive symptoms. In addition, further research is
required to further elucidate the relationship between these variables and the
mechanisms involved.
121
REFERENCES
Abramson, L.Y., Garber, J. & Seligman, M.E.P. (1980). Learned helplessness
in humans: An attributional analysis. In J. Garber & M.E.P.
Seligman (Eds.) Human helplessness: Theory and applications.
NY: Academic Press.
Aiken, L.S. & West, S.G. (1991). Multiple Regression. Newbury Park,
CA: Sage Publishers.
Alexander, P.C. (1985). A systems theory conceptualization o f incest.
Family Process, 24, 79-88.
Alexander, P.C. (1992). Application of attachment theory to the study o f
sexual abuse. Journal of Consulting and Clinical Psychology, 60,
185-195.
Allers, C.T., Benjack, K.J. & Allers, N.T. (1992). Unresolved childhood
sexual abuse: Are older adults affected? Journal of Counseling and
Development, 71, 14-17.
Allers, C.T., Benjack, K.J., White, J., & Rousey, J.T. (1993). HIV
vulnerability and the adult survivor o f childhood sexual abuse. Child
Abuse and Neglect, 17, 291-298.
Anderson , J., Martin, J., Mullen, P., Romans, S. & Herbison, P. (1993).
Prevalence o f childhood sexual abuse experiences in a community
sample o f women. Journal of the American Academy of Child and
Adolescent Psychiatry, 32, 911-919.
122
Bagley, C. (1991). The prevalence and mental health sequels o f child sexual
abuse in a community sample o f women aged 18 to 27. Canadian
Journal of Community Mental Health, 10, 103-116.
Baron, R.M. & Kenny, D.A. (1986). The moderator- mediator variable
distinction in social psychological research: Conceptual, strategic,
and statistical considerations. Journal of Personality and Social
Psychology, 51, 1173-1182.
Bartholow, B.N., Doll, L.S., Joy, D., Douglas, J.M. Harrison, J.S., Moss,
P.M. & McKiman, D. (1994). Emotional, behavioral, and HIV risks
associated with sexual abuse among adult homosexual and bisexual
men. Child Abuse and Neglect, 18, 747-761.
Biddle, B.J. & Marlin, M.M. (1987). Causality, confirmation, credulity, and
structural equation modeling. Child Development, 58, 4-17.
Bifulco, A., Brown, G.W., & Adler, Z. (1991). Extrly sexual abuse and
clinical depression in adult life. British Journal of Psychiatry, 159,
115-122.
Bifulco, A., Brown, G.W. & Harris, T.D. (1994). Childhood Experience o f
Care and Abuse (CECA): A retrospective interview measure.
Journal of Child Psychology and Psychiatry, 35, 1419-1435.
Billings, A.G. & Moos, R.H. (1982). Family environment and adaptation: A
clinically applicable typology. American Journal of Family
Therapy, 10, 26-38.
Birt, C.M. & Dion, K.L. (1987). Relative deprivation theory and responses
to discrimination in a gay male and lesbian sample. British Journal
of Social Psychology, 26, 139-145.
Blake-White, J. & Klein, C.M. (1985). Treating the dissociative processes
in adult victims o f childhood incest. Social Casework: The Journal
of Contemporary Social Work, 66, 394-402.
Bowman, E.S. (1993). Etiology and clinical course o f pseudoseizures:
Relationship to trauma, depression and dissociation. Psychosomatics,
34, 333-342.
123
Boyd, C.J. (1993). The antecedents o f women's crack cocaine abuse: family
substance abuse, sexual abuse, depression and illicit drug use.
Journal of Substance Abuse Treatment, 10, 433-438.
Bradford, J., Ryan, C., & Rothblum, E.D. (1994). National Lesbian Health
Care Survey: Implications for mental health care. Journal of
Consulting and Clinical Psychology, 62, 228-242.
Briere, J. (1992a). Child abuse traum a: Theory and treatm ent of the
lasting effects. Newbury Park, CA: Sage Publications.
Briere, J. (1992b). Methodological issues in the study o f sexual abuse effects.
Journal of Consulting and Clinical Psychology, 60, 196-203.
Briere, J. & Conte, J. (1993). Self-reported amnesia fo r abuse in adults
molested as children. Journal of Traumatic Stress, 6, 21-31.
Browne, A. & Finkelhor, D. (1986). Impact o f child sexual abuse: A review
o f the research. Psychological Bulletin, 99, 66-77.
Burbash, D.J., Kashani, J.H. & Rosenberg, T.K. (1989). Parental bonding
and depressive disorders in adolescents, Journal of Child
Psychology and Psychiatry, 30, 417-429.
Bushnell, J.A., Wells, J.E. & Oakley-Browne, M.A. (1992). Long-term
effects o f intrafamilial sexual abuse in childhood. Acta Paychiatrica
Scandinavica, 85, 136-142.
Cass, V.C. (1979). Homosexual identity formation: A theoretical model.
Journal of Homosexuality, 4, 219-235.
Cass, V.C. (1983-1984). Homosexual identity: A concept in need o f
definition. Journal of Homosexuality, 9, 105-126.
Celano, M.P. (1992). A developmental model o f victims’ internal
attributions o f responsibility for sexual abuse. Journal of
Interpersonal Violence, 7, 57-69.
Cerezo, M.A. & Frias, D. (1994). Emotional and cognitive adjustment in
abused children. Child Abuse and Neglect, 18, 923-932.
124
Christie, D. & Young, M. (1986). Self-concept o f lesbian and heterosexual
women. Psychological Reports, 59, 1279-1282.
Cliff, N. (1983). Some cautions concerning the application o f causal
modeling methods. Multivariate Behavioral Research, 18, 115-125.
Cohen, J. & Cohen, P. (1983). Applied Multiple Regression/Correlation
Analysis for the Behavioral Sciences, 2nd edition. Hillsdale, NJ:
Lawrence Erlbaum Associates.
Cole, P.M. & Putnam, F.W., (1992). Effect o f incest on self and social
functioning: A developmental psychopathology perspective. Journal
of Consulting and Clinical Psychology, 60, 174-184.
Conte, J R. & Schuerman, J R. (1987). The effects o f sexual abuse on
children: A multidimensional view. Journal of Interpersonal
Violence, 2, 380-390.
Courtois, C. (1988). Healing the incest wound: Adult survivors in
therapy. NY: W.W. Norton and Co.
Cunningham, R.M., Stiffman, A.R., Dore, P. & Earls, F. (1994). The
association o f physical and sexual abuse with HIV risk behaviors in
adolescence and young adulthood: Implications for public health.
Child Abuse and Neglect, 18, 233-245.
Cutler, S.E. & Nolen-Hoeksema, S. (1991). Accounting for sex differences
in depression through female victimization: Childhood sexual abuse.
Sex Roles, 24, 425-438.
DeBellis, M.D., Chrousos, G.P., Dorn, L.D., Burke, L., Helmers, K., Kling,
M.A., Trickett, P.K. & Putnam, F.W. (1994).
Hypothalamic-pituitary-adrenal axis dysregulation in sexually abused
girls. Journal of Clinical Endocrinology and Metabolism, 78,
249-255.
Devins, G.M. & Orme, C.M. (1985). Center for Epidemiologic Studies
Depression Scale. In D.J. Keyser & R.C. Sweetland (Eds.), Test
Critiques, Vol. n (pp. 144-160). Kansas City, MO: Test
Corporation of America.
125
DeYoung, M. & Lowry, J.A. (1992). Traumatic bonding: Clinical
implications in incest. Child Welfare, 71, 165-175.
Donaldson, M.A. & Cordes-Green, S. (1994). Group treatment of adult
incest survivors. Newbury Park, CA: Sage Publications.
Drossman, D.A., Leserman, J., Nachman, G., Zhiming, L., Gluck, H.,
Toomey, T.C. & Mitchell, M. (1990). Sexual and physical abuse in
women with functional or organic gastrointestinal disorders. Annals
of Internal Medicine, 113, 828-833.
Dutton, D.G. & Painter, S. (1993). Emotional attachments in abusive
relationships: A test o f traumatic bonding theory. Violence and
Victims, 8, 105-120.
Edwards, J.J. & Alexander, P.C. (1992). The contribution of family
background to the long-term adjustment o f women abused as children.
Journal of Interpersonal Violence, 7, 306-320.
Ernst, C., Angst, J. & Foldenyi, M. (1993). XVII. Sexual abuse in
childhood. Frequency and relevance for adult morbidity data o f a
longitudinal epidemiological study. European Archives of
Psychiatry and Clinical Neuroscience, 242, 293-300.
Erwin, K. (1993). Interpreting the evidence: Competing paradigms and the
emergence o f lesbian and gay suicide as a "social fact."
International Journal of Health Services, 23, 437-453.
Finkelhor, D. & Browne, A. (1985). The traumatic impact of child sexual
abuse: A conceptualization. American Journal of Orthopsychiatry,
55, 530-541.
Finkelhor, D., Araji, S., Baron, L., Browne, A., Peters, S.D. & Wyatt, G.E.
(1986). A sourcebook on child sexual abuse. Newbury Park, CA:
Sage Publications.
Finkelhor, D. (1990). Early and long-term effects o f child sexual abuse: An
update. Professional Psychology: Research and Practice, 21,
325-330.
126
Finkelhor, D. & Dziuba-Leatherman, J (1994). Victimization o f children.
American Psychologist, 49, 173-183.
Fox, K.M. & Gilbert, B.O. (1994). The interpersonal and psychological
functioning o f women who experienced childhood physical abuse,
incest and parental alcoholism. Child Abuse and Neglect, 18, 849-
858.
Garnefski, N., van Egmond, M., & Straatman, A. (1990). The influence o f
early and recent life stress on severity o f depression. Acta
Psychiatrica Scandinavica, 81, 295-301.
Gatz, M. & Karel, M. (1990). Positive and negative mood factors on the
CES-D over time. Paper presented at the meeting of the American
Psychological Association, Boston, MA.
Gelinas, D.J. (1983). The persisting negative effects o f incest. Psychiatry,
46, 312-332.
Green, A.H. (1988). Child maltreatment and its victims: A comparison o f
physical and sexual abuse. Psychiatric Clinics of North America,
11 (4), 591-610.
Gidycz, C.A., Coble, C.N., Latham, L. & Layman, M.J. (1993). Sexual
assault experience in adulthood and prior vicimization experiences.
Psychology of Women Quarterly, 17, 151-168.
Goddard, C.R. & Stanley, J.R. (1994). Viewing the abusive parent and the
abused child as captor and hostage: The application o f hostage
theory to the effects o f child abuse. Journal of Interpersonal
Violence, 9, 258-269.
Gold, E.R. (1986). Long term effects o f sexual victimization in childhood:
An attributional approach. Journal of Consulting and Clinical
Psychology, 54, 471-475.
Grand, S. & Alpert, J.L. (1993). The core trauma o f incest: An object
relations view. Professional Psychology: Research and Practice,
24, 330-334.
127
Gurian, A. (1987). Depression and young girls: Early sorrows and
depressive disorders. In. R. Formanek & A. Gurian (Eds.), Women
and depression: A lifespan perspective. NY: Springer Publishing
Company.
Harter, S., Alexander, P.C. & Neimeyer, R.A. (1988). Long-term effects o f
incestuous child abuse in college women: Social adjustment, social
cognition, and family characteristics. Journal of Consulting and
Clinical Psychology, 56, 5-8.
Hartman, C.R. & Burgess, A.W. (1993). Information processing o f trauma.
Child Abuse and Neglect, 17, 47-58.
Henry, J.P. (1993). Psychological and physiological responses to stress: The
right hemisphere and the hypothalamus-pituitary-adrenal axis, an
inquiry into the problems o f human bonding. Integrative
Physiological and Behavioral Science, 28, 369-387.
Herman, J.L. (1981). Father-Daughter incest. Cambridge, MA: Harvard
University Press.
Herman, J.L. (1992). Trauma and recovery. NY: Basic Books.
Herman, J.L., Perry, C. & Van der Kolk, B. (1989). Childhood trauma in
borderline personality disorder. American Journal of Psychiatry,
146, 490-495.
Herman, J.L., Russell, D. & Trocki, K. (1986). Long-term effects o f
incestuous abuse in childhood. American Journal of Psychiatry,
143, 1293-1296.
Herman, J.L. & Schatzow, E. (1987). Recovery and verification o f memories
o f childhood sexual trauma. Psychoanalytic Psychology, 4, 1-14.
Herman, J.L. & Van der Kolk, B.A. (1987). Traumatic antecedents o f
borderline personality disorder. In B.A. Van der Kolk (Ed.),
Psychological Trauma. Washington, DC: American Psychiatric
Press.
128
Hsu, K. & Marshall, V. (1987). Prevalence o f depression and distress in a
large sample o f Canadian residents, interns, and fellows. American
Journal of Psychiatry, 144, 1561-1566.
Jack, D.C. (1991). Silencing the self: Women and depression.
Cambridge, MA: Harvard University Press.
Jackson, S.W. (1986). Melancholia and depression: From Hippocratic
times to modern times. New Haven, CT: Yale University Press.
Janus, S.S. & Janus, C.L. (1993). The Janus Report on Sexual Behavior.
New York: John Wiley and Sons.
Jehu, D. (1988). Beyond sexual abuse: Therapy with women who were
childhood victims. Chichester, UK: John Wiley and Sons.
Jehu, D. (1989). Sexual dysfunctions among women clients who were
sexually abused in childhood. Behavioral Psychotherapy, 17, 53-70.
Kaplan, A. (1964). The conduct of inquiry: Methodology for behavioral
science. Scranton, PA: Chandler.
Kaplan, H.B. & Pokorny, A.D. (1969). Self-derogation and psychosocial
adjustment. Journal of Nervous and Mental Diseases, 149, 421-434.
Kendall-Tackett, K.A., Williams, L.M. & Finkelhor, D. (1993). Impact o f
sexual abuse on children: A review and synthesis o f recent empirical
studies. Psychological Bulletin, 113, 164-180.
Kendler, K.S., Kessler, R.C., Neale, M.C., Heath, A.C. & Eaves, L.J. (1993).
The prediction o f Major Depression in women: Toward an integrated
etiologic model. American Journal of Psychiatry, 150, 1139-1148.
Kenny, D.A. (1979). Correlation and causality. New York: John Wiley
and Sons.
Kluft, R.P. (Ed.). (1985). Childhood antecedents of multiple personality.
Washington, D.C.: American Psychiatric Press.
Koverola, C., Pound, J., Heger, A. & Lytle, C. (1993). Relationship o f child
sexual abuse to depression. Child Abuse and Neglect, 17, 393-400.
129
Lerner, H.G. (1987). Female depression: Self-sacrifice and self-betrayal in
relationships. In. R. Formanek & A. Gurian (Eds ), Women and
depression: A lifespan perspective. NY. Springer Publishing
Company.
Loftus, E., Garry, M. & Feldman, J. (1994). Forgetting sexual trauma:
What does it mean when 38% forget? Journal of Consulting and
Clinical Psychology, 62, 1177-1181.
Long, P.J. & Jackson, J.L. (1994). Childhood sexual abuse: An examination
o f family functioning. Journal of Interpersonal Violence, 9,
270-277.
Longstreth, G.F. & Wolde-Tsadik, G. (1993). Irritable bowel-type symptoms
in HMO examinees. Digestive Diseases and Sciences, 38, 1581-1589.
Martin, D. & Hetrick, E.S. (1988). The stigmatization o f the gay and
lesbian adolescent. Journal of Homosexuality, 15,163-183.
McCann, I.L. & Pearlman, L.A. (1990). Psychological trauma and the
adult survivor: Theory, therapy and transformation. NY:
Brunner/Mazel.
McGrath, E., Keita, G.P., Strickland, B.R. & Russo, N.F. (Eds.). (1990).
Women and depression: Risk factors and treatment issues.
Washington, D.C.: American Psychological Association.
Meiselman, K.C. (1978). Incest: A psychological study of causes and
effects with treatment. San Francisco: Jossey Bass.
Minton, H.L. & McDonald, G.J. (1983-1984). Homosexual identity
formation as a developmental process. Journal of Homosexuality, 9,
91-104.
Muenzenmaier, K., Meyer, I., Struening, E. & Ferber, J. (1993). Childhood
abuse and neglect among women outpatients with chronic mental
illness. Hospital and Community Psychiatry, 44, 666-670.
130
Murrey, G.J., Bolen, J., Miller, N., & Simensted, K. (1993). History o f
childhood sexual abuse in women with depressive and anxiety
disorders: A comparative study. Sex Education and Therapy, 19,
13-19.
Nash, M.R., Hulsey, T.L., Sexton, M.C., Harralson, T.L. & Lambert, W.
(1993). Long-term sequelae o f childhood sexual abuse: Perceived
family environment, psychopathology, and dissociation. Journal of
Consulting and Clinical Psychology, 61, 276-283.
Neumann, D.A. (1991). A Self Psychology Model for the Long-term
Effects of Childhood Sexual Abuse. Unpublished doctoral
dissertation, Fuller Theological Seminary, Pasadena, CA.
Neumann, D.A. (1994). Long-term correlates o f childhood sexual abuse in
adult survivors. In J. Briere (Ed ), Assessing and treating victims of
violence. New Directions for Mental Health Services, Number 64.
San Francisco. Jossy-Bass Publishers.
O’Brien, E.J. (1985). Global self-esteem scales: Unidimensional or
multidimensional? Psychological Reports, 57, 383-389.
Orme, J., Reis, J. & Herz, E.J. (1986). Factorial and discriminant validity
o f the Center fo r Epidemiological Studies Depression (CES-D) scale.
Journal of Clinical Psychology, 42, 28-33.
Padesky, C.A. (1989). Attaining and maintaining positive lesbian self
identity: A cognitive therapy approach. Women and Therapy, 8,
145-156.
Paris, J., Zweig-Frank, H. & Guzden, H. (1993). The role o f psychological
risk factors in recovery from borderline personality disorder.
Comprehensive Psychiatry, 34, 410-413.
Parker, G. (1983). Parental "affectionless control" as an antecedent to adult
depression: A risk factor delineated. Archives of General
Psychiatry, 40, 956-960.
Parker, G. (1989). Parental Bonding Instrument: Psychometric Properties
Reviewed. Psychiatric Developments, 4,317-335.
131
Parker, G., Tupling, H., & Brown, L.B. (1979). A Parental Bonding
Instrument. British Journal of Medical Psychology, 52, 1-10.
Peters, S.D. (1988). Child sexual abuse and later psychological problems.
In G.E. Wyatt & G.J. Powell (Eds.), Lasting effects of child sexual
abuse (pp. 101-118). Newbury Park, CA: Sage Publications.
Pribor, E.F. & Dinwiddie, S.H. (1992). Psychiatric correlates o f incest in
childhood. American Journal of Psychiatry, 149, 52-56.
Putnam, F.W., Guroff, J.J., Silberman, E.K., Barban, L. & Post, R.M.
(1986). The clinical phenomenology o f multiple personality disorder:
A review o f 100 recent cases. Journal of Clinical Psychiatry, 47,
285-293.
Roesler, T.A. & McKenzie, N. (1994). Effects o f childhood trauma on
psychological functioning in adults sexually abused as children.
Journal of Nervous and Mental Disease, 182, 145-150.
Rosenberg, M. (1965). Society and the adolescent self-image. Princeton,
NJ. Princeton University Press.
Rosenthal, R. & Rosnow, R.L. (1991). Essentials of behavioral research:
Methods and data analysis. NY: McGraw-Hill, Inc.
Rosenthal, R.H. (1993). Psychology o f chronic pelvic pain. Obstetrics and
Gynecology Clinics of North America, 20, 627-642.
Rothblum, E.D. (1990). Depression among lesbians: An invisible unresolved
phenomenon. Journal of Gay and Lesbian Psychotherapy, 1, 67-87.
Rothblum, E.D. (1994). "/ only read about myself on bathroom walls": The
need for research on the mental health o f lesbians and gay men.
Journal of Consulting and Clinical Psychology, 62, 221-227.
Russell, D. (1986). The Secret Trauma. NY: Basic Books.
Schetky, D.H. (1990). A review o f the literature on the long-term effects o f
childhood sexual abuse. In R.P Kluft (Ed.), Incest-related
syndromes of adult psychopathology. Washington, DC: American
Psychiatric Press.
132
Shahani, C., Dipboye, R.L. & Phillips, A.P. (1990). Global self-esteem as a
correlate o f work-related attitudes: A question o f dimensionality.
Journal of Personality Assessment, 54, 276-288.
Sophie, J. (1987). Internalized homophobia and lesbian identity. Journal
of Homosexuality, 14, 53-65.
Stein, J.A., Golding, J.M., Siegel, M.& Burnam, M. (1988). Long-term
psychological sequelae o f child sexual abuse: The Los Angeles
epidemiologic catchment area study. In G.E. Wyatt and G.J. Powell
(Eds.), Lasting effects of child sexual abuse (pp. 135-154).
Newbury Park, CA: Sage Publications.
Summit, R. (1983). The child sexual abuse accomodation syndrome. Child
Abuse and Neglect, 7, 177-193.
Terr, L. (1988). What happens to early memories o f trauma? A study o f
twenty children under age five at the time o f documented traumatic
events. Journal of the American Academy of Child and
Adolescent Psychiatry, 27, 96-104.
Terr, L. (1990). Too Scared to Cry. NY: Basic Books.
Terr, L. (1991). Childhood traumas: An outline and overview. American
Journal of Psychiatry, 148, 10-20.
Troiden, R.R. (1988) Homosexual identity development. Journal of
Adolescent Health Care, 9, 105-113.
Tsai, M. & Wagner, N. (1978). Therapy groups for women sexually
molested as children. Archives of Sexual Behavior, 7, 417-427.
Valentine, L. & Feinauer, L.L. (1993). Resilience factors associated with
female survivors o f childhood sexual abuse. American Journal of
Family Therapy, 21, 216-224.
Van der Kolk, B. (1989). The compulsion to repeat the trauma:
Re-enactment, revicimization, and masochism. Psychiatric Clinics of
North America, 12, 389-411.
133
Van der Kolk, B., Perry, C. & Herman, J.L. (1991). Childhood origins of
self-destructive behavior. American Journal of Psychiatry, 148,
1665-1671.
Wakefield, H. & Underwager, R. (1992). Recovered memories of alleged
sexual abuse: Lawsuits against parents. Behavioral Sciences and
the Law, 10, 483-507.
Walker, E.A., Katon, W.J., Hansom, J., Harrop-Griffiths, J, Holm, L., Jones,
M.L., Hickok, L. & Jamelka, R.P. (1992). Medical and psychiatric
symptoms in women with childhood sexual abuse. Psychosomatic
Medicine, 54, 658-664.
White, M.T. & Weiner, M.B. (1986). The Theory and Practice of Self
Psychology. NY: Brunner/Mazel Publishers.
Williams, L.M. (1994a). Recall of childhood trauma: A prospective study
of women’ s memories o f child sexual abuse. Journal of Consulting
and Clinical Psychology, 62, 1167-1176.
Williams, L.M. (1994b). What does it mean to forget child sexual abuse?:
A reply to Loftus, Garry, and Feldman (1994). Journal of
Consulting and Clinical Psychology, 62, 1182-1186.
Wind, T.W. & Silvern, L. (1992). Type and extent o f child abuse as
predictors of adult functioning. Journal of Family Violence, 7, 261-
281.
Wind, T.W. & Silvern, L. (1994). Parenting and family stress as mediators
of the long-term effects o f child abuse. Child Abuse and Neglect, 18,
439-453.
Wyatt, G.E. (1985). The sexual abuse of Afro-American and White-American
women in childhood. Child Abuse and Neglect, 9, 507-519.
Wyatt, G.E. & Newcomb, M.D. (1990). Internal and external mediators of
women’ s sexual abuse in childhood. Journal of Consulting and
Clinical Psychology, 58, 758-676.
134
Wyatt, G.E., Newcomb, M.D., Riederle, M.H. (1993). Sexual abuse and
consensual sex: Women’s developmental patterns and outcomes.
Newbury Park, CA: Sage Publications.
Yama, M.F., Tovey, S.L. & Fogas, B.S. (1993). Childhood family
environment and sexual abuse as predictors o f anxiety and depression
in adult women. American Journal of Orthopsychiatry, 63,
136-141.
Zweig-Frank, H. & Paris, J. (1991). Parent’ s emotional neglect and
overprotection according to the recollections o f patients with
Borderline Personality Disorder. American Journal of Psychiatry,
148, 648-651.
135
APPENDIX
The following is the text of the ad placed in Sojourner
and Lesbian News.
LESBIAN RESEARCH PROJECT
I am a Lesbian PhD candidate doing dissertation research designed to describe the experiences of lesbian
women.
Your Help is Needed
If you are willing to spend 20-30 minutes filling out my survey, please fill in the form below and send it to
Chris Diana Cooper, MA
Dept of Psychology (SGM-501)
University of Southern California
Los Angeles, CA 90089-1061
My research is directed at issues around depression, self-esteem and childhood sexual abuse. HOWEVER, I
need to hear from you as a lesbian woman whether these issues affect you directly or not- I am
especially interested in hearing from lesbian women of color as well as women from all class and
educational backgrounds. ALL RESPONSES WILL BE ANONYMOUS. I will destroy records of your
name and address upon sending research materials. Psychological and medical research ignores us. You can
help by being part of one study that asks about lesbian women.
NAME ^ —
ADDRESS _____________________________________________________
CITY _____________________________________ STATE ZIP
137
The following is a list of the questions that make up the Parental Bonding
Instrument. The statements in conventional type comprise the "caring" subscales
( m o t h e r -c a r in g and f a t h e r -c a r in g ). Statements in italics comprise the
"overprotection" subscales ( m o t h e r -o v e r p r o t e c t iv e and f a t h e r -o v e r p r o t e c t iv e ).
Statements preceded by an "R" have coding reversed before adding total scores
for subscales, so high scores on the "caring" subscales reflect greater perceived
caring or warmth and high scores on "overprotection" subscales reflect greater
intrusiveness and control.
R Spoke to me with a warm and friendly voice
Did not help me as much as I needed
Let me do those things I liked doing
Seemed emotionally cold to me
R Appeared to understand my problems and worries
R Was affectionate to me
Liked me to make my own decisions
R Did not want me to grow up
R Tried to control everything I did
R Invaded my privacy
R Enjoyed talking things over with me
R Frequently smiled at me
R Tended to baby me
Did not seem to understand what I needed or wanted
Let me decide things for myself
Made me feel I wasn’t wanted
R Could make me feel better when I was upset
Did not talk to me very much
R Tried to make me dependent on her/him
R Felt I could not look after myself unless she/he was around
Gave me as much freedom as I wanted
Let me go out as often as I wanted
R Was overprotective o f me
Did not praise me
Let me dress the way L pleased
138
SURVEY FORM
The following pages contain a copy of the cover-letter and survey form used in
this research.
139
October 1, 1993
Dear Participant,
Enclosed is a questionnaire designed to help me with the data collection for my PhD dissertation
in Clinical Psychology at the University of Southern California. I am asking for your help in
completing these forms and returning them to me. I would like to emphasize that you, of
course, may at any point, choose NOT to participate. I know you will give it careful
consideration and I appreciate your time and efforts.
As you know, there is very little psychological research done in the lesbian community. Most
research is done in the heterosexual world and then generalized to us. Clearly, what is true of
heterosexual women may not apply to lesbian women. It is because of my belief that lesbian
women need representation in the research, that 1 have begun this project which is designed to
investigate a number of psychological issues, and includes questions about depression, self
esteem, relationships with parents, and sexual experiences.
One of the enclosed questionnaires concerns experiences in your childhood that involve touching
or sexual activity. Please be aware that if in your comments you specifically identify (by name)
a person who is currently abusing a child (or children), I am required to report that information
to the state Department of Social Services which might then initiate an investigation.
If you have any questions or wish to discuss this, please feel free to call me (Chris Diana
Cooper, MA) at (818) 308-7906.
Again, many thanks for your time and your help.
Sincerely,
PLEASE FILL OUT INFORMATION ON THE FRONT AND BACK OF THE
NEXT FOUR PAGES. ALL QUESTIONNAIRES SHOULD BE RETURNED IN
THE ENCLOSED ENVELOPE.
140
FOR THE FOLLOWING ITEMS, PLEASE INDICATE THE EXTENT TO WHICH YOU AGREE
OR DISAGREE:
STRO NGLY
AGREE AGREE DISAGREE
I feel that I’m a person of w orth,
at least on an equal basis w ith others
I feel that I have a number o f good
qualities
A ll in all, I am inclined to feel that
I am a failure
I am able to do things as well as
m o st other people
I feel I do not have m uch to b e proud of
I take a positive attitude tow ard m yself
O n the whole, I am satisfied w ith m yself
I w ish I could have more respect for
m y self
I certainly feel useless at times
At tim es I think I am no good at all
STRONGLY
DISAGREE
4
4
4
4
4
4
4
4
4
141
FOR THE FOLLOWING ITEMS, PLEASE INDICATE THE EXTENT TO WHICH YOU AGREE
OR DISAGREE:
D U R IN G THE PA ST W EEK
I w as bothered by things that usually
d o n ’t bother me
I did not feel like eating, my appetite
w as poor
I felt that I could not shake off the
b lu es even w ith help from my family or
friends
I felt that I was ju st as good as other
people
I had trouble keeping my mind on w hat
I w a s doing
I felt depressed
I felt that everything I did was an effort
I felt hopeful about the future
I thought my life had been a failure
I felt fearful
M y sleep was restless
I w as happy
I talked less than usual
I felt lonely
People were unfriendly
I enjoyed life
I h ad crying spells
I felt sad
I felt that people disliked me
I could not "get going"
STRONGLY
AGREE
STRONGLY
DISAGREE DISAGREE
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
142
This questionnaire lists various attitudes and behaviors o f parents. As you rem em ber your m other and
father in your first 16 years w ould you circle the most appropriate num ber after each question ( 0 =
not at all true and 3 = very true)
"M other and Father" may not describe your situation. (For exam ple, you m ay have been bom to and
raised by two lesbian m others, or you may have been raised by a birth m other and a step-father. If
you have and can rem em ber a father and mother, please till this out in relation to them. If some
other circum stances apply, please amend the appropriate headings and fill out the form accordingly.
S p o k e to m e w ith a w a rm a n d frie n d ly v o ic e 0
MOTHER
1 2 3 0
FATHER
1 2 3
D id n o t h e lp m e a s m o d i a s I n e e d e d 0 1 2 3 0 1 2 3
L et m e d o th o e e th in g s I lik e d d o in g 0 1 2 3 0 1 2 3
S eem ed e m o tio n a lly c o ld to m e 0 1 2 3 0 1 2 3
A p p e are d to u n d e n ta n d m y p ro b le m s a n d w o rrie s 0 1 2 3 0 1 2 3
W as a ffe c tio n a te to m e 0 1 2 3 0 1 2 3
i.iifrf m e to m a k e m y o w n d e c isio n s 0 1 2 3 0 1 2 3
D id n o t w a n t m e to g ro w u p 0 1 2 3 0 1 2 3
T ried to c o n tro l e v e ry th in g I d id 0 1 2 3 0 1 2 3
In v a d e d m y p riv a c y 0 1 2 3 0 1 2 3
E njoyed ta lk in g th in g s o v « w ith m e 0 1 2 3 0 1 2 3
F re q u e n tly s m ile d a t m e 0 1 2 3 0 1 2 3
T en d e d to b a b y m e 0 1 2 3 0 1 2 3
D id n o t se e m to u n d e rs ta n d w h a t I n e ed e d o r w a n ted 0 1 2 3 0 1 2 3
L et m e d e rid e th in g s fo r m y a e if 0 1 2 3 0 1 2 3
M ad e m e fe e l I w a s n 't w a rd e d 0 1 2 3 0 1 2 3
C o u ld m a k e m e fe e l b e lts ’ w h e n 1 w a s u p se t 0 1 2 3 0 1 2 3
D id n o t ta lk w ith m e v e ry m u c h 0 1 2 3 0 1 2 3
T rie d to m a k e m e d e p e n d e n t o n h i m / h s 0 1 2 3 0 1 2 3
F rit I c o u ld n 't lo o k a fte r m y s e lf u n le ss (s)h e w a s a ro u n d 0 1 2 3 0 1 2 3
G a v e m e a s m u c h fre e d o m a s I w a n ted 0 1 2 3 0 1 2 3
L et m e g o o u t a s o f t s i a s I w a n te d 0 1 2 3 0 1 2 3
W as o v e rp ro te c tiv e o f m e 0 1 2 3 0 1 2 3
D id n o t p ra is e m e 0 1 2 3 0 1 2 3
L et m e d ie a e in a n y w a y I p le a s e d 0 1 2 3 0 1 2 3
Have either o f your parents (or primary caretakers while
growing up) ever had treatm ent for psychological problems? YES NO UNSURE
If yes, who? (circle as m any as apply) M OTHER FA TH ER OTHER____________
If yes, did treatm ent include hospitalization? YES NO UNSURE
D o either o f your parents (or prim ary caretakers while
growing up) drink alcohol? YES NO UNSURE
If yes, w ould you say they used alcohol to excess? YES NO UNSURE
143
EARLY SEXUAL EXPERIENCES
Q uestion 1
Before you were age 17, did anyone ever kiss you or touch your
body (but not on your genital organs) in a sexual w ay?
If yes:
a) Did this happen with som eone 5 or m ore years older than
you w ere?
b) About how many tim es w ere you kissed or touched non-genitally
in a sexual way by som eone 5 or more years older before age 17?
c) Who kissed you or touched you non-genitally in a sexual way
before the age of 17? (please check all that apply)
a friend
a date
a stranger
a family member ( fa th e r; m o th e r; sibling)
a friend of the family
a babysitter or nanny
a teacher, doctor, church official or other professional
som eone not m entioned above
d) Did anyone ever use physical force on these occasions?
e) Did anyone ever use verbal threats or pressure on these
occasions?
Q uestion 2
Before you were age 17, did anyone ever touch your sexual parts
in a sexual way, or did you touch their sexual parts? YES NO
If yes:
a) Did this happen with som eone 5 or more years older than
you w ere? YES NO
b) About how many times did this type o f touching occur w ith
someone 5 or more years older before age 17? _______
c) Who touched you genitally (or whom did you touch) before
the age o f 17? (please ch eck all that apply)
a friend
a date
a stranger
a fam ily member ( fa th e r; m o th e r; sibling)
a friend o f the family
a babysitter or nanny
a teacher, doctor, church official or other professional
som eone not m entioned above
YES NO
YES N O
YES NO
YES N O
UNSURE
UNSURE
UNSURE
UNSURE
UNSURE
UNSURE
144
Q uestion 2 (Continued)
d) Did anyone ever use physical force on these occasions?
e) Did anyone ever use verbal threats or pressure on
these occasions?
Q uestion 3
Before you w ere age 17, did anyone ever have vaginal, oral,
or anal intercourse w ith you, or place their finger or
objects in your anus or vagina?
If yes:
a) Did this happen with som eone 5 or more years older
than you were?
b) A bout how m any tim es did someone 5 or more years older
than you have vaginal, oral or anal intercourse with
you before age 17?
c) W ho had vaginal, oral or anal intercourse w ith you before
the age o f 17? (please check all that apply)
a friend
a date
a stranger
a family m em ber ( fa th e r; m o th e r; sibling)
a friend o f the family
a babysitter or nanny
a teacher, doctor, church official or other professional
som eone not m entioned above
d) D id anyone ever use physical force on these occasions?
e) Did anyone ever use verbal threats or pressure on
these occasions?
YES N O UNSURE
YES N O UNSURE
YES N O UNSURE
YES N O UNSURE
YES N O UNSURE
YES N O UNSURE
IF YOUR ANSWERS TO QUESTIONS 1. 2. AND 3 WERE "NO." SKIP TO QUESTION 6
Q uestion 4
Regarding any experiences described in Question 1, Question 2, and/or Question 3:
a) At the tim e o f the experience(s), were you able to talk
to anyone about the experience(s)?
If yes, did you feel that you were listened to and believed?
b) Have you EVER talked w ith a friend or family mem ber about
this experience?
If yes, about how long w as it between the first time
o f the experience and w hen you talked about it with a
friend or family mem ber?
YES N O UNSURE
YES N O UNSURE
YES N O UNSURE
145
Question 4 (Continued^
c) Have you ever talked with a psychotherapist, counsellor
or other professional about this experience? YES
If yes, about how long was it betw een the first tim e of
the experience and when you talked about it w ith a
psychotherapist, counselor or o th er professional?________________________ __
Q uestion 5
Regarding any experiences described in Question 1, Question 2,
and/or Question 3:
a) To the best o f your knowledge have you always remembered
these experiences? YES
If no, did you (at any time) forget all o f the
experience(s), or only part of it or them?
b) Do you think you currently rem em ber all of what
happened to you? YES
Q uestion 6
To the best o f your knowledge, would you say that you were
sexually abused as a child (before age 17)? YES
Q uestion 7
W ould you say that you w ere physically ab used as a child
(before age 17)? YES
Q uestion 8
Overall, how would you rate your childhood:
Very unhappy Average V ery happy
1 2 3 4 5 6 7
Com m ents foptionali
N O UN SURE
N O UN SURE
A LL PART
N O U N SU RE
N O UN SURE
N O U N SU R E
146
1. Do you now considered yourself to be depressed? YES NO
2. Do you believe that you have been depressed in the past? Y E S NO
IF YOU ANSWERED "YES" TO EITHER QUESTION 1 OR 2 ABOVE:
Do you think your depression is/was connected to
or caused by your sexual orientation Y E S NO
In w hat way?
Do you think your depression is/was connected to
or caused by childhood sexual abuse? Y E S NO
In w hat way?
W hat do you think is the origin of y o u r sexual orientation (genetic, childhood experiences,
upbringing, etc). Please explain as fully as you care to.
I f you wish to receive a report o f the findings o f this research project, please fill out the
enclosed postcard and mail it separately from the questionnaires.
147

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Childhood precursors of adult reality distortion: A prospective study.

PDF

A correlational study of feminist/womanist identity development and depression in women

PDF

Biosocial antecedents of schizophrenia-spectrum personality disorders: A longitudinal study

PDF

An interpretive communication study of images and roles of women in selected situation comedies from 1950-1975

PDF

An examination of reliable component analysis

PDF

A cross-cultural comparison of marital power and dyadic adjustment among American, Indo-American, and East Indian dual-career and single-career couples

PDF

Concordant and discordant drug use in intimate relationships: A longitudinal study

PDF

Childhood videotaped neuromotor and social precursors of schizophrenia: A prospective investigation

PDF

An evaluation of a brief HIV /AIDS prevention intervention using normative feedback to promote risk reduction among sexually active college students

PDF

Child sexual abuse in a sample of male and female Hispanic and White nonclinical adolescents: Extending the reliability and validity of the Trauma Symptom Inventory (TSI)

PDF

A stress model of acculturation, depressive symptoms, and self-esteem among Mexican-American undergraduates

PDF

Attachment as a predictor of self -monitoring and sexual behavior

PDF

Adolescent predictors and adult consequences of alcohol dysphoria: A longitudinal study

PDF

Behavioral and demographic predictors of breast cancer stage at diagnosis

PDF

Affective Information And Affect-Cognition Interaction: Mood And Emotion

PDF

Client characteristics, service characteristics, and psychosocial outcomes in the community treatment of schizophrenia

PDF

Experimental evaluation of a new intervention designed to promote dementia caregivers' acceptance and empathic responding towards care -recipients

PDF

"In censure of his seeming": External marking, fashion and travel in the English Renaissance.

PDF

Characteristics of college males who are likely to sexually harass women: A test of a mediated model.

PDF

Expectancies for alternative behaviors predict drinking problems: A test of a cognitive reformulation of the matching law

Asset Metadata

Creator Cooper, Christine Diana (author)

Core Title Childhood Sexual Abuse and Depressive Symptoms In A Lesbian Population: An Exploratory Study.

Degree Doctor of Philosophy

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag OAI-PMH Harvest,psychology, clinical,women's studies

Language English

Contributor Digitized by ProQuest (provenance)

Advisor Earleywine, Mitchell (committee chair), [illegible] (committee member), Farver, Joann (committee member)

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c17-73610

Unique identifier UC11354748

Identifier 9625011.pdf (filename),usctheses-c17-73610 (legacy record id)

Legacy Identifier 9625011.pdf

Dmrecord 73610

Document Type Dissertation

Rights Cooper, Christine Diana

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the au...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus, Los Angeles, California 90089, USA