J

. Child Psychol. Psychiat. Vol. 40, No. 3, pp. 465–477, 1999

Cambridge University Press

' 1999 Association for Child Psychology and Psychiatry

Printed in Great Britain. All rights reserved

0021–9630

\99 $15.00j0.00

Avoidant

\Ambivalent Attachment Style as a Mediator between Abusive

Childhood Experiences and Adult Relationship Difficulties

Gerard McCarthy and Alan Taylor

Medical Research Council’s Child Psychiatry Unit, Institute of Psychiatry, London, U.K.

The role of attachment style, self-esteem, and relationship attributions as possible mediators
between abusive childhood experiences and difficulties in establishing supportive love
relationships in adulthood were investigated in a sample of women known to be at risk of
experiencing relationship problems. Measures of child abuse, the quality of love relation-
ships, and the three potential mediators were made concurrently in adulthood. Participants
who had experienced child abuse were found to be six times more likely to be experiencing
difficulties in the domain of adult love relationships than those who had not. Self-esteem and
relationship attributions were not found to be related to child abuse. When both child abuse
and avoidant

\ambivalent attachment style were considered together avoidant\ambivalent

attachment style, but not child abuse, was found to be related to relationship difficulties.
These findings indicate that avoidant

\ambivalent attachment style, but not self-esteem and

relationship attributions, is a mediating factor in the route from child abuse to adult
relationship abilities.

Keywords :

Child abuse, adult love relationships, mediators, avoidant

\ambivalent attach-

ment style.

Abbreviations :

A

\C: avoidant-resistant; APFA: Adult Personality Functioning Assess-

ment ; BDI : Beck Depression Inventory ; CSA : child sexual abuse ; ICC : interclass
correlation ; RAM : Relationship Attribution Measure ; RSD : Rosenberg self-derogation
scale ; SESS : Self-Evaluation and Social Support Instrument.

Introduction

A number of studies have shown that abusive child-

hood experiences are related to difficulties in establishing
supportive cohabiting relationships in adulthood. Studies
into the long-term outcome of childhood sexual abuse
have found increased rates of interpersonal difficulties
and sexual problems (Beitchman, Zucker, & Hood, 1992 ;
Finkelhor, 1983 ; Finkelhor, Hotaling, Lewis, & Smith,
1990 ; Mullen, Martin, Anderson, Romans, & Herbison,
1994). Children who have experienced physical abuse or
harsh and

\or neglectful parenting have also been found

to experience difficulties in the domain of intimate adult
relationships (Andrews & Brown, 1988 ; Birtchnell, 1993 ;
Brown & Moran, 1994 ; Malinosky-Rummell & Hansen,
1993 ; Quinton, Pickles, Maughan, & Rutter, 1993). These
findings are important because research has established
that difficulties in close relationships, such as marital
discord and lack of support, are related both to the
development of emotional and behavioural problems in
the next generation (Emery, 1982 ; Farrington, 1995 ;

Requests for reprints to : Dr Gerard McCarthy, Child and
Adolescent Service, Weston Clinic, 4a Beaconsfield Rd,
Weston-super-Mare, BS23 1YE, U.K.

Jouriles, Piffner, & O’Leary, 1988 ; Patterson, 1982) and
to the development of mental health problems in adult
life (Brown & Harris, 1978 ; Quinton & Rutter, 1988 ;
Sampson & Laub, 1993).

As the evidence for a link between abusive childhood

experiences and adult relationship difficulties grows so
the need to understand how this effect occurs becomes
more urgent. A major issue for all concerned with the
well-being of children and adolescents is how to intervene
to prevent them going on to experience difficulties in
adulthood. However, little is known about the psycho-
logical processes underlying this association nor about
the role cognitive-affective mechanisms may play in
mediating the effects of abusive early experiences. A
better understanding of the precise mechanisms through
which negative childhood experiences exert their effect on
later psychosocial functioning is likely to have impli-
cations for methods of intervention to prevent the
development of later problems.

At present, attempts to delineate the role of specific

processes in mediating developmental continuities be-
tween abusive childhood experiences and adulthood are
hampered by the nature of the evidence available. Much
of this evidence comes from either short-term longi-
tudinal studies in childhood or from cross-sectional
retrospective studies undertaken in adulthood (Maughan
& McCarthy, 1997). Researchers investigating the link
between child abuse and adult depression have recently

465

466

G. M

CARTHY and A. TAYLOR

begun to undertake direct tests of the role of specific
mediating factors. This work has shown that stable
negative characteristics of the self may mediate the
association between child abuse and adult depression. In
particular two mediating factors have been identified.
These are bodily shame (Andrews, 1995) and charactero-
logical self-blame (Andrews & Brewin, 1990 ; Gold, 1986).
More recently Andrews (1997) has shown that bodily
shame also acts as a mediator between child abuse and
bulimia in a community sample of young women.

The present study attempts to build on this small body

of research to identify specific psychological factors that
mediate links between child abuse and adverse psycho-
logical functioning in adulthood. However, it differs in
two principal ways from the studies by Andrews and her
colleagues. First, rather than focusing on a particular
psychiatric outcome such as depression or bulimia, we
aimed to investigate psychological processes that mediate
the link between child abuse and the quality of adult love
relationships. As noted above, adult relationship abilities
are hypothesised to be one of the pathways by which
adverse early experiences lead to increased risk of
psychopathology in adult life. Second, rather than testing
the role of a single mediating factor such as self-esteem,
this study aimed to investigate the role of a number of
potential mediators. In particular, attachment style,
relationship attributions, and self-esteem were investi-
gated as possible mediating factors between abusive
experiences in childhood and later difficulties in the
domain of adult love relationships in a sample of adult
women.

Attachment Style

According to Bowlby’s theory of attachment (Bowlby,

1969, 1973, 1980) children develop cognitive

\affective

representations or internal working models of their
experiences in their attachment relationships. These are
hypothesised to provide a model about how close
relationships typically proceed and they are thought to
influence the quality of later love relationships (Bowlby,
1979, 1988 ; Crowell & Treboux, 1995). Indeed, a central
tenet of attachment theory is ‘‘ that parent–child relation-
ships are prototypes of later love relationships ’’ (Crowell
& Treboux, 1995, p. 296). Until recently, internal working
models have mainly been investigated in the context of
infants’ relationships with their principal caregivers,
where they are inferred on the basis of infant’s behaviour
in separation-reunion procedures (Ainsworth, Blehar,
Waters, & Wall, 1978). New methods have now been
developed to assess older children’s internal working
models of attachment (Main, Kaplan, & Cassidy, 1985)
and adult’s representations of their childhood experiences
with parents (Main, 1991). At the same time, social and
clinical psychologists have begun to investigate the role of
working models of attachment or attachment styles
in adult romantic relationships (Collins & Read, 1990 ;
Hazan & Shaver, 1987). However, although there are
growing literatures on both the relationship between
maltreatment and the development of insecure attach-
ment relationships in childhood and the relationship
between attachment styles and the quality of love
relationships in adulthood, little is known about the role

internal working models may play in mediating continu-
ities in relationship functioning between childhood and
adulthood.

Investigations of attachment in maltreated children

have shown high levels of insecure (avoidant and am-
bivalent) attachments (Crittenden, 1988 ; Egeland &
Sroufe, 1981 ; Schneider-Rosen, Braunwald, Carlson, &
Cicchetti, 1985). In an inner-city disadvantaged sample,
Egeland and Sroufe (1981) found a specific association
between child abuse and the development of avoidant
attachment although subsequent investigations have not
replicated this finding. More recent research has shown
high levels of atypical attachment patterns that do not fit
smoothly into Ainsworth et al.’s (1978) traditional A-B-C
classification scheme (Main & Soloman, 1986). Critten-
den (1988) found that most maltreated children in her
sample could be classified as having avoidant-ambivalent
(A

\C) patterns of attachment and Carlson, Cicchetti,

Barnett, and Braunwald (1989) found that over 80 % of
maltreated infants had disorganized

\disoriented attach-

ments (Main & Soloman, 1986). Lynch and Cicchetti
(1991) have demonstrated that distortions in maltreated
children’s mental representations of attachment may
persist into the preadolescent years. They found that
30 % of maltreated children between the ages of 7 and 13
years reported having confused patterns of relatedness to
their mothers.

Recent work on adult love relationships has shown

that adult attachment style is associated with the quality
of these relationships (Hazan & Shaver, 1987 ; Shaver &
Hazan, 1993). Studies have demonstrated that attach-
ment style is related to levels of satisfaction, commitment,
and love and trust in the relationship (Collins & Reed,
1990 ; Davis, Kirkpatrick, Levy, & O’Hearn, 1994 ;
Mikulincer & Erev, 1991 ; Simpson, 1990) and with
observed levels of self-disclosure and emotional support
(Mikulincer & Nachshon, 1991 ; Simpson, Rholes, &
Nelligan, 1992). Secure attachment style is associated
with a desire for intimate relationships, and within
relationships secure adults seek a balance of closeness
and autonomy and are comfortable with feelings of
dependency. Anxious-ambivalent attachment style is
characterised by a desire for close relationships, but
fear of rejection may lead ambivalent adults to seek
extreme forms of intimacy and lower levels of autonomy.
Avoidant attachment style is associated with a need
to maintain distance and avoidant adults are thought
to feel uncomfortable with feelings of intimacy and
dependency (Shaver & Hazan, 1993). Bartholomew
(1990 ; Bartholomew & Horowitz, 1991) has further
differentiated the avoidant style into a fearful avoidant
pattern and a dismissing avoidant pattern. Although
both types share a negative view of others, dismissing
avoidants have a positive view of self, whereas fearful
avoidants also have a negative view of self. Fearful
avoidant attachment style is known to be related to
a number of high-risk environments such as having
problem drinking parents (Brennan, Shaver, & Tobey,
1991) and being the victim of incestuous abuse
(Alexander, 1993). A number of studies have also
demonstrated that retrospective accounts of childhood
relationships with parents are related in theoretically
meaningful ways with adult attachment styles (Shaver &

467

AVOIDANT

\AMBIVALENT ATTACHMENT STYLE

Hazan, 1993). However, at present empirical investi-
gations into the role attachment representations may
play in mediating developmental continuities between
childhood adversity and the quality of adult love
relationships have yet to be undertaken.

Relationship Attributions

There are two main reasons for thinking that attribu-

tional processes may play a role in mediating continuities
between childhood experiences and later relationship
difficulties. First, research by Dodge and his colleagues
has demonstrated that children who have been physically
abused and harshly parented tend to acquire deviant
patterns of processing social information and that these
predict the development of later aggressive behaviour
(Dodge, Bates, & Pettit, 1990 ; Dodge, Pettit, Bates, &
Valente, 1995 ; Weiss, Dodge, Bates, & Pettit, 1992). In
particular, physically abused children acquired a bias to
over-attribute hostile intent to others’ behaviour. Conse-
quently, children who have experienced physical abuse in
childhood may be at increased risk of developing mal-
adaptive attributional processes that, if carried into later
close relationships, may lead to difficulties in this domain.
For example, children who have experienced harsh
parenting and deliberate acts of physical or emotional
cruelty may be at increased risk of interpreting future
negative partner behaviour in close relationships as being
intentionally motivated, and therefore deserving of blame
and retaliation.

Second, a number of studies have demonstrated that

the attributions that partners make for relationship events
are related to marital satisfaction (Baucom, 1987 ; Brad-
bury & Fincham, 1990). Distressed partners have been
found to make causal and responsibility attributions that
accentuate the impact of negative events. Causal attri-
butions concern the explanations a partner makes for the
occurrence of a relationship event and responsibility
attributions deal with the accountability or answerability
for the event (Bradbury & Fincham, 1990). Researchers
have found that the attributions partners make for
relationship events are related to behaviour in adult love
relationships (Bradbury, Beach, Fincham, & Nelson,
1996 ; Bradbury & Fincham, 1992). Husbands who are
physically violent to their wives have also been shown to
be more likely than nonviolent husbands to attribute
negative intentions, selfish motivation, and blame to the
wife (Holtzworth-Munroe, 1992 ; Holtzworth-Munroe &
Anglin, 1991 ; Holtzworth-Monroe & Hutchinson, 1993).

Whereas current maladaptive attributional processes

may be a product of being in an unhappy current
relationship, it is also possible that maladaptive attri-
butional processes play a causal role in bringing about
problems in intimate relationships (Bradbury & Fincham,
1992). Indeed, maladaptive attributional processes may
pre-date the present relationship altogether and may have
their origins in an earlier developmental period. Recently,
researchers have begun to investigate associations be-
tween child abuse and maladaptive attributional pro-
cesses in adulthood. Gold (1986) found evidence of
greater dispositional blame for hypothetical bad events in
people who reported having been sexually victimised in
childhood. More recently, Andrews (1992) found that

retrospective reports of repeated physical and sexual
abuse in childhood were associated with a tendency for
women in violent marriages to blame their own characters
for the marital violence. Following on from these studies
we aimed to investigate the role causal and responsibility
attributions may play in mediating links between child
abuse and difficulties in adult love relationships.

Self

-esteem

Many theorists have argued that difficulties in making

close adult relationships derive from negative thoughts
and feelings about the self (Erikson, 1963 ; Kohut, 1977 ;
Sullivan, 1953). The idea that low self-esteem may
mediate developmental continuities between adverse
childhood experiences and difficulties in adulthood comes
from a number of areas.

First, research shows that negative childhood experi-

ences are related to impairments in various aspects of
children’s sense of self. Maltreated toddlers have been
shown to have more negative responses to visual self-
recognition experiments, and to experience more diffi-
culties in dealing with issues to do with autonomy and
exploration, than comparison children (Egeland &
Sroufe, 1981 ; Schneider-Rosen & Cicchetti, 1984). Older
maltreated children describe themselves as being less
competent and accepted than comparison children and
appear to have feelings of generalised low self-worth
(Vondra, Barnett, & Cicchetti, 1989). Also, in a series of
studies, Sroufe and his colleagues have shown that the
quality of care children receive in their early relationships
with parents is related to a number of measures of
children’s subsequent sense of self (Sroufe, 1986, 1989).
Children who received more sensitive care as infants tend
to have higher self-esteem, to have greater ego strength,
to be more ego resilient, to be more independent and
resourceful, and to have more elaborate and complex
fantasy play than children who received less sensitive care
(Arend, Gove, & Sroufe, 1979 ; Egeland & Sroufe, 1981 ;
Matas, Arend, & Sroufe, 1978 ; D. Rosenberg, 1984 ;
Sroufe, 1983 ; Waters, Wippman, & Sroufe, 1979).

Second, whereas these studies show how negative early

experiences can adversely affect the development of self-
esteem in childhood, a series of studies by Brown and his
colleagues suggest that these processes may be involved in
mediating continuities between lack of affectionate care
in childhood, negativity in current close relationships,
and vulnerability to depression (Andrews & Brown, 1988,
1993 ; Brown, Bifulco, & Andrews, 1990 ; Brown, Bifulco,
Veiel, & Andrews, 1990 ; Brown & Moran, 1994). Also of
importance here is evidence showing that the protective
effects of successful coping and positive experiences may
exert their effect on the developmental process by
bringing about an increase in self-confidence and self-
esteem (Rutter, 1989).

The Nature of Mediating Variables

Given the lack of research in this area the main aim of

this study was to carry out a preliminary investigation
into the role of these three potential mediating factors in
the link between abusive childhood experiences and later
problems in making supportive love relationships. An

468

G. M

CARTHY and A. TAYLOR

Figure 1.

Mediator model.

important limitation of the present study was that
measures of the three potential mediators were made in
adulthood rather than in childhood, concurrently with
measures of functioning in the domain of adult love
relationships. Ideally, prospective longitudinal studies
are required to provide a more powerful test of potential
mediating variables. However, until suitable prospec-
tively studied samples are available, much can be learned
from retrospective accounts of childhood experiences.
Prospective research in this area is also costly and time
consuming. In light of this it seemed reasonable to use the
present study to explore as far as possible potential
mediators that might profitably be incorporated in future
prospective work.

In order to test the value of our hypothesised medi-

ators, we used the framework outlined by Baron and
Kenny (1986) to guide our analyses. According to Baron
and Kenny a variable can be thought of as functioning as
a mediator if it meets the following conditions (see Fig.
1) : (1) if variations in levels of the independent variable,
in this case child abuse, significantly account for vari-
ations in the presumed mediator (Path A) ; (2) if variations
in the proposed mediator significantly account for vari-
ations in the dependent variable, in this case relationship
quality (Path B) ; (3) when paths (A) and (B) are
controlled, a previously significant relation between the
independent variable (child abuse) and the dependent
variable (relationship quality) is no longer significant.
From a theoretical perspective, when path C is reduced to
zero this provides strong evidence for a single dominant
mediator, whereas a significant but less complete re-
duction in the residual path C demonstrates that a given
variable plays an important role in mediating an effect.

Research has also shown that abusive childhood

experiences are often associated with a range of other
negative childhood experiences such as lack of parental
warmth and involvement and poor family functioning
(Bifulco, Brown, & Adler, 1991 ; Fergusson, Lynskey, &
Horwood, 1996 ; Mullen et al., 1993). This raises the
possibility that the apparent correlation between child
abuse and adverse psychosocial functioning in adult-
hood may account for the background factors that pre-
dispose to child abuse and not the abuse itself (Mullen
et al., 1993 ; Paradise, Rose, Sleeper, & Nathanson, 1994).
Indeed, there are a number of reports that raise the
possibility that poor family functioning may account for
much of the variance in the outcomes of child sexual
abuse victims (Fromouth, 1986 ; Harter, Alexander, &
Neimeyer, 1998 ; Wyatt & Mickey, 1987). Recently,
Mullen and his colleagues attempted to disentangle the

particular effects of child sexual abuse (CSA) from a
number of other negative childhood experiences (Mullen
et al., 1993). They found that while a number of other
factors also related to psychosocial difficulties in adult-
hood, CSA continued to have a direct negative impact
when these factors were controlled for. Results from the
Christchurch Health and Development Study have con-
firmed this finding (Fergusson, Horwood, & Lynskey,
1996). In order to address these issues a high-risk group
of subjects was used, all of whom had experienced poor
parenting in childhood. This enabled us to investigate
specific processes that may contribute to the impact of
child abuse on adult relationship difficulties.

The main aim of the study was therefore to investigate

whether each of the potential mediators met the con-
ditions outlined for judging whether a variable can be
said to function as a mediator. In particular the study
aimed to test whether child abuse was significantly related
to the three mediating variables (path A), whether the
mediating variables were significantly related to the
quality of adult love relationships (path B), and whether,
when paths A and B were controlled, a previously
significant relationship between child abuse and relation-
ship problems became nonsignificant. Such a situation
would provide evidence that a variable functions as a
mediator between child abuse and adult relationship
problems.

Method

Participants

Forty women known to have experienced poor parenting in

childhood took part in the study. The sample was mainly drawn
from a number of previous longitudinal follow-up studies of
children living in inner-city London (Champion, Goodall, &
Rutter, 1995 ; Maughan & Hagell, 1996 ; Quinton et al., 1993).
During these studies participants had been interviewed about
their childhood experiences and this information was used to
establish a sample of women all of whom had experienced poor
parent–child relationships. A total of 44 women were identified
who met the criterion for having experienced poor parenting in
childhood, described in more detail below, but we were not able
to trace the current whereabouts of 5 of them. The remainder
(N

l 39) were contacted by letter to explain the purpose of the

new study and to ask whether they would agree to participate
and 5 (13 %) women refused to take part. The 34 women who
agreed to take part were interviewed in their homes. We felt it
was necessary to try to supplement the sample in order to
increase the number of cases available for statistical analyses. In
order to do this a further sample of women were recruited from
an inner-city general practice in South London. A total of 18
women from the general practice were interviewed and 6 were
found to meet the criteria for inclusion in the study outlined
below.

The average age of the women was 35

n7 and ages ranged from

25 to 42 years. Twenty-nine of the women were currently in a
cohabitation of at least 6 months duration ; 11 were not
currently cohabiting. Twenty-one of the women were currently
married and 10 were divorced or separated.

Measures

Childhood experiences

. Adults were interviewed about their

childhood experiences using a nonscheduled standardised in-
terview (Brown & Rutter, 1966). The interview included detailed
assessments of the quality of parent–child relationships. All of

469

AVOIDANT

\AMBIVALENT ATTACHMENT STYLE

the women reported having experienced a poor relationship
with one of their parents. This was defined as having experienced
any two of the following three indicators in childhood or
adolescence : (1) little or no warmth, (2) harsh or lax discipline,
(3) little or no communication or involvement in activities.

The women were also questioned in detail about abusive

experiences in childhood. Abuse in childhood was defined as
that occurring before the age of 16 years.

(1) Sexual abuse was defined as that involving direct physical

contact of the sexual parts. Willing sexual contact with
nonrelated peers in the teenage years was excluded. These
criteria are similar to those used in other studies that have
examined the relationship between childhood sexual
abuse and adult psychosocial functioning (Andrews,
1995 ; Andrews & Brown, 1988 ; Bifulco, Brown, &
Harris, 1994).

(2) Physical abuse was defined as repeated acts of violence by

a parent towards the child and included acts such as being
punched, burnt, kicked, hit with an instrument, or being
tied up.

(3) Emotional abuse referred to situations where the

parent(s) expressed marked negative feelings toward the
child. This included expressions of hostility or scape-
goating of the child. The hostile behaviour was required
to be persistent over time and personally focused on the
child.

Attachment style

. Attachment style was assessed using the

Hazan and Shaver (1987) adult attachment questionnaire. This
is a standard procedure for measuring adult attachment style.
In the first part subjects were given the three Hazan and Shaver
descriptions of attachment styles and asked to rate on a 7-point
scale (1

l disagree strongly and 7 l agree strongly), how much

they agreed or disagreed with each style as a description of the
way they generally are in love relationships. For example, the
secure prototype reads as follows : ‘‘ I find it relatively easy to get
close to others and am comfortable depending on them. I don’t
often worry about being abandoned or about someone getting
too close to me ’’. The avoidant prototype reads : ‘‘ I am some-
what uncomfortable being close to others ; I find it difficult to
trust them completely, difficult to allow myself to depend on
them. I am nervous when anyone gets too close, and often, love
partners want me to be more intimate than I feel comfortable
being ’’. The ambivalent prototype reads ; ‘‘ I find that others are
reluctant to get as close as I would like. I often worry that my
partner doesn’t really love me or want to stay with me. I want
to get very close to my partner, and this some times scares
people away ’’. In the second part subjects were asked to choose
the attachment style that best described the way they typically
feel in romantic relationships. In our analyses we have focused
on the attachment scales rather than the discrete attachment
categories, to capitalise on the advantages of dimensional
analysis (Collins & Read, 1990). First, a limitation of discrete
measures is that inevitably some members represent the
category ‘‘ better ’’ than others. Second, dimensional scales can
help to retain information lost through a forced classification
procedure, especially where individuals report high scores on
two or more of the attachment scales. Third, the use of scale
measures is also useful in terms of statistical power when sample
sizes are small, as in our study (Fergusson & Horwood, 1995).
In particular we have focused on the avoidant and ambivalent
scales, as these were most likely to be involved in mediating
developmental continuities between child abuse and adult
relationship abilities.

Self

-esteem. This was assessed using the Rosenberg self-

esteem scale (M. Rosenberg, 1965). This is a 10-item Gutman
scale tapping global positive and negative attitudes to the self. It
has been used with a wide variety of subject populations and

shows good test–retest reliability (r

l n85). There is evidence

suggesting that this scale reflects two independent aspects of
self-esteem, tapped by the positively and negatively worded
items (Kaplan & Pokorny, 1969 ; Kohn & Schooler,
1983 ; Zeller & Carmaines, 1980) and that only the negative
measure is linked to childhood adversity and to negative
psychosocial outcomes (Andrews & Brown, 1993 ; Kaplan
& Pokorny, 1969). Therefore only the score based on the
negative items (range 5 to 20) was used in the following
analyses. Following Kaplan and Pokorny (1969) and Andrews
and Brown (1993) we refer to the negative measure as the
Rosenberg self-derogation (RSD) scale.

Attributions

. The procedure used in this study was a modi-

fication of the Relationship Attribution Measure (RAM)
developed by Fincham and Bradbury (1992). This measure has
high internal consistency and test–retest reliability. In the RAM
subjects are asked to imagine six hypothetical negative partner
behaviours and after each one they are asked to rate their
agreement with six statements on a 6-point scale. Subjects make
the ratings after imagining that the behaviour has just occurred
in their relationship. Each scale point is labelled ranging from
disagree strongly to agree strongly. The first three statements
are used to assess three different types of causal attributions.
These are the extent to which the cause rested with the partner
(locus), is likely to change (stability), and affects other areas of
the relationship (globality). The second three statements are
used to assess partner intent, motivation, and blame.

In the procedure used here subjects were asked to think about

the most important love relationships they had been involved
in. Then each scenario was read out by the interviewer and the
subject was asked to imagine themselves in the hypothetical
situation. After this the subject was asked to rate their
agreement with the six attribution statements from the RAM.
The stories were used to depict the following hypothetical
negative partner behaviours : jealousy, negative mood, lack of
interest, lack of support in carrying out household chores, an
unexpected let down, and abandonment. Many of these themes
had been used in previous research in the field (Fincham &
Bradbury, 1992 ; Holtzworth-Munroe & Hutchinson, 1993).
Hypothetical behaviours were used because they enable stan-
dard stimuli to be presented to all participating subjects and
because they could be rated by subjects who were not currently
in a love relationship. Following the RAM, negative partner
events were used because previous research has shown that
attributions for negative events are related more consistently
and more strongly to marital satisfaction than are attributions
for positive events (Baucom, Epstein, Sayers, & Goldman-Sher,
1989 ; Fincham, Beach, & Nelson, 1987).

Examples of the situations include the following :

(1) A woman and her partner have two young children and

she notices that he has not completed the chores and tasks
that he said he would do around the house.

(2) A woman wants to tell her partner something that has

been on her mind for a while and is important to her.
Finally she tells him and as she does she notices that he is
not paying attention to her and he is watching the
television.

Composite attribution indices were formed for responses across
the 18 causal attribution responses (3 dimensions

i6 stimulus

events) and the 18 responsibility attribution responses (3
dimensions

i6 stimulus events). Higher scores on each subscale

reflected attributions that accentuated the impact of the negative
partner behaviour (e.g. see it as more stable, intentional, and
blameworthy). These composite attribution indices were found
to be reliable. The alpha coefficient for the causal attribution
composite was (

α l n86) and for the responsibility attribution

composite it was (

α l n93).

470

G. M

CARTHY and A. TAYLOR

Interview measures of love relationships

. The quality of

subjects’ love relationships was assessed using the appropriate
section from the Adult Personality Functioning Assessment
(APFA), (Hill, Harrington, Fudge, Rutter, & Pickles, 1989).
The APFA assesses functioning in a range of domains (work,
love relationships, friendships, everyday coping, nonspecific
social contacts, and negotiations with officials and others).
Unlike trait-based measures of personality functioning, the
APFA depends on detailed descriptions of behaviour and
ratings are made by the interviewer from relatively extended
periods (of 5 years at the least). In this study ratings were made
for the previous 5 years. Questioning in each domain covers a
range of issues. In relation to love relationships, questioning
covers relationship histories during the rating period, and the
extent of shared activities, support, confiding, arguments, and
violence in each relevant relationship. The interview also
provides evidence of any deviance (defined as criminality,
drinking, problems of drug abuse) in participants’ partners’
functioning. Each domain is rated on a 6-point scale, with a
rating of 1 indicating exceptionally positive functioning, 2
indicating good functioning, 3 reflecting relatively minor or
transient problems, and ratings of 4, 5, and 6 indicating
increasing degrees of difficulty. For example, in love relation-
ships if a person has been married or has cohabited for several
years, and there have not been any major difficulties, a rating of
1 or 2 is made. When the subject has had persistent discord
leading to repeated breakdown of relationships, or when all
relationships have lacked confiding and support, or there was
an absence of sustained committed relationships, a rating of 5
or 6 is made. A rating of 4, 5, or 6 was taken as indicating that
a person had experienced difficulties in the domain of adult love
relationships over the last 5 years and a rating of 1, 2, or 3 was
taken as indicating positive functioning in this domain. The
APFA shows good inter-rater reliability with interclass cor-
relations (ICCs) ranging from

n65 to n88 in individual domains

and a mean ICC of

n77 (Hill et al., 1989). The ICC for the love

relationships domain was

n81.

Beck Depression Inventory

(BDI ) (Beck, 1984). An additional

aim of the study was to investigate the impact of current mood
on the pattern of associations predicted above. For example,
current mood states may influence reports of the past or they
may lead to distortions or biases in the reporting of more recent
events or experiences (Lewinsohn & Rosenbaum, 1987). Con-
sequently, associations between child abuse, relationship
difficulties, and the potential mediators could in part be
accounted for by current mood state. In order to investigate
this issue the BDI was used. This is a widely used 21-item
questionnaire designed to assess the severity of current de-
pressive symptoms.

Statistical Methods

The majority of analyses were carried out using SPSS (SPSS

Inc., 1994). Due to the small sample size of our study the
association between categorical variables was tested using
Fisher’s Exact test. Differences between groups on the con-
tinuous measures were compared using Mann-Whitney U tests
using exact p values as the scales were non-normal. Correlations
between scales were investigated using Spearman’s rank coef-
ficient. For the final mediation models exact logistic regression
using the package LogXact (LogXact, 1996 ; Mehta & Patel,
1995) was used. This enabled us to avoid problems with tests
based on large sample approximations that are used in standard
maximum likelihood logistic regression (Mehta & Patel, 1995).
All analyses were carried out on 39 cases except for the analyses
that included the BDI and the self-esteem scale, each of which
had one case with missing data.

Results

Child Abuse and Adult Relationships : Descriptive
Statistics

The incidence of child abuse was high in our sample.

Forty-one per cent (16

\39) of women with complete data

reported having experienced some form of child abuse :
six reported having been sexually abused, seven reported
having been physically abused, and three reported emo-
tional abuse. Three women reported having experienced
more than one form of child abuse. As expected, a high
level of negative functioning in adult love relationships
were found in this sample. Forty-nine per cent (19

\39)

were rated as having negative functioning in the domain
of adult love relationships over a 5-year period (rating of
4, 5, or 6 on the APFA) whereas 51 % (20

\39) were rated

as having positive functioning in this domain (rating of 1,
2, or 3 on the APFA).

Association between child abuse and poor adult relation

-

ships

. Child abuse and negative functioning in adult love

relationships were strongly associated. Among those who
had experienced child abuse 75 % (12

\16) were rated as

having experienced difficulties in adult love relationships
compared with only 30 % (7

\23) of those who had not

experienced child abuse (Fisher’s Exact Test, p

l n01).

Effects of current mood

. We investigated the possible

effect of current mood on the association between child
abuse and negative functioning in adult love relationships
and found that the BDI measure was not significantly
related to child abuse (Table 1). However, the level of
current depressive symptoms was considerably higher in
the group with poor adult relationships compared to the
group with good relationships, this difference just failing
to reach significance. As depressed mood was not
associated with child abuse it cannot act as a mediator of
the association between child abuse and poor adult
relationships. Therefore, the BDI measure has not been
included in subsequent analyses.

Possible Mediators : Descriptive Statistics

Theoretically we were interested in three possible

mediators of the association between child abuse and
difficulties in the domain of adult love relationships : self-
esteem, attachment style, and attributional style. De-
scriptive statistics for measures of these characteristics
are as follows : (1) Self-esteem : RSD scale mean 13

n3 (SD

3

n6); (2) Relationship attribution scales: responsibility

scale mean 64

n1 (SD 17n7), causal scale mean 67n8 (SD

12

n9); (3) Attachment scales: avoidant scale mean 3n4

(SD 2

n2), ambivalent scale mean 3n3 (SD 2n2). Although

we have not used the secure attachment scale and the
forced-choice attachment classification we include them
here for completeness. As expected, in our sample there
was a high rate of insecure attachment styles, with 41 %
of women rating themselves as having an avoidant
attachment style (16

\39) and 15% (6\39) rating them-

selves as anxious-ambivalent. The remaining 44 % rated
themselves as having a secure attachment style (17

\39).

This distribution of subjects across the three attachment

471

AVOIDANT

\AMBIVALENT ATTACHMENT STYLE

Table 1
Association between Measures

Experienced child abuse

Poor adult relationships

Yes

No

Test

Yes

No

Test

Variable

Mean

(SD)

Mean

(SD)

statistic

a

p

Mean

(SD)

Mean

(SD)

statistic

a

p

Beck Depression Inventory

9

n5

(9

n2)

8

n8

(6

n8)

170

n0

n953

11

n6

(8

n9)

6

n8

(5

n7)

116

n5

n063

Self-esteem

12

n5

(3

n2)

13

n8

(3

n8)

128

n0

n191

12

n1

(3

n2)

14

n4

(3

n6)

101

n0

n021

Attribution scales

Responsibility

71

n5 (16n6)

59

n0 (17n0)

122

n5

n079

73

n9 (13n9)

54

n9 (16n1)

68

n5

n001

Causal

72

n8 (11n1)

64

n3 (13n2)

121

n0

n074

74

n2

(9

n7)

61

n8 (12n8)

61

n0

n001

Avoidant attachment scale

4

n4

(2

n2)

2

n7

(2

n0)

109

n5

n032

4

n7

(2

n1)

2

n2

(1

n6)

66

n0

n001

Ambivalent attachment scale

4

n2

(2

n4)

2

n7

(1

n8)

116

n5

n053

4

n4

(2

n1)

2

n3

(1

n7)

80

n0

n002

Avoidant

\ambivalent scale

8

n6

(3

n9)

5

n4

(3

n2)

98

n5

n014

9

n1

(3

n3)

4

n4

(2

n6)

53

n5

n001

Secure attachment scale

3

n3

(2

n1)

4

n5

(2

n1)

115

n5

n050

3

n1

(1

n8)

4

n9

(2

n1)

90

n0

n004

a

All tests based on Mann–Whitney U.

categories was different from those found in other studies.
In their 1987 study Hazan and Shaver, for example,
found that 55 % were secure, 25 % avoidant, and 20 %
ambivalent, whereas Feeney and Noller (1990), in their
study of Australian students, found 55 % were secure,
30 % avoidant, and 15 % ambivalent. In particular, lower
rates of secure attachment style and higher rates of
avoidant attachment style were found here. The mean of
the secure attachment scale was 4

n0 (SD l 2n1).

The pattern of associations between the three proposed

mediating constructs (attachment style, relationship attri-
butions, and self-esteem) were also investigated, and
significant correlations were found between the measures.
The relationship attribution measures were correlated
with the two attachment scales, particularly the avoidant
scale, which had a correlation of

n43 (p l n006) with the

causal scale and

n42 (p l n008) with the responsibility

scale (correlations for the ambivalent scale were : .37,
p

l n019 and n35, p l n029, respectively). Scores on the

RSD were negatively correlated with both of the attach-
ment scales and the attributional scales. The individual
correlations were as follows : avoidance scale

lkn40,

p

l n013; ambivalence scale lkn56, p

n001; causal

scale

l kn39, p l n015; and responsibility scale lkn33,

p

l n042. The avoidant and ambivalent attachment

scales were positively correlated (correlation

l n51,

p

l n001), as were the attributional style scales (corre-

lation

l n58, p n001).

Possible Mediators : Preliminary Analyses

As a first stage in our analyses we looked at the relation

between our potential mediators and the experience of
child abuse to determine the magnitude and significance
of their association and therefore the possibility of
mediation. Results of these analyses are given in Table 1.
The measures of self-esteem and attributional style were
not significantly related to the experience of child abuse in
this sample and therefore could not act as mediators to
problems in adult functioning. For the RSD, the mean
scores by experience of child abuse were slightly lower in
the group who had experienced child abuse. The attri-
butional style measures gave more differentiation be-
tween the two groups, but the difference between groups

was above significance levels. Child abuse was signifi-
cantly related to attachment style with the difference
between groups being largest for the avoidance scale,
whereas the difference between groups for the ambiv-
alence scale was of slightly smaller magnitude and just
over the nominal significance level (see Table 1). As the
above analyses indicated that both types of insecure
attachment were related to child abuse and were highly
correlated, we created a further scale by summing the
individual ambivalence and avoidance scales. This re-
sulted in an overall avoidant

\ambivalent scale, with high

values indicating an attachment style characterised by
high level of both ambivalence and avoidance. We
considered it preferable to use this measure rather than
the secure attachment measure because it indexes negative
attachment process explicitly and in further analyses we
will focus primarily on this attachment measure. Based
on this measure, those who had experienced child abuse
had a much higher score on the avoidant

\ambivalent

scale than those who had not experienced child abuse
(see Table 1).

Attachment style and poor adult love relationships

. In

order to function as a mediator between poor adult
relationships and the experience of child abuse the
attachment measure also needed to be associated with
relationship difficulties. When we investigated this as-
sociation we found a strong link with avoidant

\

ambivalent attachment scale and difficulties in the do-
main of adult love relationships. The mean of this scale
among those with good functioning in this domain was
4

n4 (SD 2n6) compare to a mean of 9n1 (SD 3n3) for those

with poor functioning.

Other mediators and poor adult love relationships

.

Although attributions were not significantly related to
child abuse, they were strongly related to poor function-
ing in the domain of adult love relationships (see Table 1).
For example, subjects with poor functioning in the
domain of adult relationships had more negative scores
on the responsibility scale. This indicated that they were
more likely to attribute negative intent and blame than
those who did not have poor adult relationships. The
RSD scale was also related to poor relationships in
adulthood, with those having experienced poor relation-
ships having significantly lower RSD scores (see Table 1).

472

G. M

CARTHY and A. TAYLOR

Table 2
Logistic Regression Models of Poor Adult Relationships

95 % CI

OR

Lower

Upper

Score

a

p

b

Model 1

Child abuse

6

n48

1

n36

38

n27

7

n31

n006

Model 2

Child abuse

3

n70

0

n51

32

n90

2

n43

n142

Avoidant

\ambivalent scale

1

n56

1

n18

2

n24

11

n17

n001

a

Exact conditional scores test (df

l 1).

b

Mid-p value.

Attachment style as a mediator between child abuse and

adult relationship difficulties

. To investigate the potential

mediating effect of attachment style on the relation
between child abuse and poor adult relationships we used
exact logistic regression models ; the results are presented
in Table 2. Model 1, with only the effect for child abuse
fitted, indicates the strength of the relation between child
abuse and poor adult relationships. Those who had
experienced child abuse were over six times more likely to
have difficulties in the domain of intimate adult relation-
ships than those who had not. The large bounds to the
confidence interval for this effect are due to the small
sample size of the study, which results in a large degree
of uncertainty about the magnitude of the effect,
particularly in its upper bound. In order to test the
possible mediating effect of attachment style on child
abuse the avoidant

\ambivalent scale was added to the

model (Model 2). This scale was strongly related to poor
adult relationships—the more insecure individuals being
more likely to have poor functioning in the domain of
adult love relationships. The effect of child abuse in
this model was markedly diminished, becoming non-
significant and with the odds ratio falling by 43 % to 3

n7,

indicating that an avoidant

\ambivalent attachment style

is a strong mediator in the route from child abuse to
poor adult relationships.

Discussion

Results from this study provide support for the idea

that the link between abusive childhood experiences and
difficulties in making supportive cohabiting relationships
in adulthood are mediated in part by an avoidant

\

ambivalent attachment style. We have been able to
demonstrate that the avoidant

\ambivalent attachment

scale meets the three conditions outlined by Baron and
Kenny (1986) for judging whether a variable can be said
to function as a mediator. First, abusive childhood
experiences were found to be significantly related to the
avoidant

\ambivalent scale and second, the avoidant\

ambivalent scale was significantly related to difficulties in
the domain of intimate adult relationships. Finally, a
previously significant relationship between child abuse
and difficulties in adult love relationships became non-
significant when the above two pathways were controlled.
Importantly, this pattern of findings could not be
accounted for by current depressive mood. Previous
research has suggested that current mood states may be
associated with distortions or biases in the reporting of

recent experiences (Lewinsohn & Rosenbaum, 1987).
Ratings of current mood on the BDI were not sig-
nificantly related to reports of having experienced child
abuse nor to adult relationship difficulties. Also of interest
was the finding that the other two potential mediating
variables, self-esteem and relationship attributions, did
not appear to play a significant role in mediating the
effects of abusive childhood experiences on later re-
lationship abilities. We have been able to establish that
child abuse appears to have a direct negative impact on
functioning in later adult love relationships in a sample
where all the women had experienced poor parenting
in childhood, and that this link is partly mediated by
attachment processes. Child abuse is often associated
with other measures of poor family functioning and it has
been suggested that the latter may account for much of
the variance in the outcomes of child abuse victims
(Fromouth, 1986 ; Harter et al., 1988 ; Wyatt & Mickey,
1987). One of the particular strengths of this study’s
design was that we were able to explore these specific links
in a sample where all the women had experienced adverse
childhood experiences.

Of particular interest was the finding that an avoidant

\

ambivalent style appeared to play an important role in
mediating the effects of child abuse on adult relationship
difficulties. This suggests that child abuse may increase
the risk of developing an attachment style that is
characterised by high levels of avoidance and ambiv-
alence. That is, abusive experiences in childhood may
increase the risk of developing multiple and contradictory
strategies for dealing with attachment-related issues in
intimate relationships and this may give rise to recurrent
oscillations between a desire for extreme forms of
intimacy and a desire to maintain distance. For example,
when moving closer to romantic partners, women who
have been abused in childhood may feel anxious about
the prospect of developing an intimate relationship and
this may lead them to distance themselves from the
partner. However, when they move away, they may feel
uncomfortable about being alone and this may in turn
fuel the desire again for more extreme forms of intimacy.
The finding that an avoidant

\ambivalent attachment

style may play a role in mediating the effects of child
abuse on adult relationship difficulties is in line with
findings from research on attachment in childhood and
adulthood, where atypical attachment categories have
been found in a range of high-risk groups (Belsky &
Cassidy, 1994). Main and Soloman (1986) identified a
fourth pattern known as disorganised

\disoriented in

473

AVOIDANT

\AMBIVALENT ATTACHMENT STYLE

children and Crittenden (1988) called these atypical
patterns avoidant-resistant (A

\C) as they are charac-

terised in part by a co-mingling of avoidant and am-
bivalent behaviour. Rates of these attachment classi-
fications are known to increase as the severity of social
risk factors increases (Carlson et al., 1989 ; Cicchetti &
Toth, 1995 ; Lyons-Ruth, Connell, Grunebaum, &
Botein, 1990), and researchers have also recently es-
tablished that disorganised attachment classification is
associated with the development of hostile and aggressive
behaviour in children both at home and in the classroom
(Lyons-Ruth, Alpern, & Repacholi, 1993). Latty-Mann
(1989, 1991), in a study of adults who had grown up in
families with an alcoholic parent, identified a subgroup of
persons who rated themselves as experiencing high levels
of avoidance and ambivalence in their love relationships.
Latty-Mann suggested that this attachment style could be
called ‘‘ Ambivalent ’’ because they appeared both to
want closeness but also to fear it and its consequences.
Latty-Mann and Davis (1996) have recently provided
evidence suggesting that this new attachment style is
closely related to Bartholomew’s Fearful-Avoidant at-
tachment style (Bartholomew, 1990). Importantly, Fear-
ful attachment style is also known to be related to a
number of high-risk environments such as having had
problem drinking parents (Brennan et al., 1991) and
being the victim of incestuous abuse (Alexander, 1993).
Interestingly, Brennan et al. suggested, following Latty-
Mann and Davis (1988), that the fearful avoidant
category in Bartholomew’s scheme may be analogous to
the disorganised or A

\C category in infancy (Crittenden,

1988 ; Main & Soloman, 1986).

The fact that the residual pathway between child abuse

and relationship difficulties was not reduced to zero in
our final analyses suggests that an avoidant

\ambivalent

attachment style is one of a number of mediating factors
rather than the single dominant mediator. More research
is required to identify other mediating factors. These may
include other cognitive-affective variables such as dys-
functional patterns of emotional regulation (Cicchetti,
Ganiban, & Barnett, 1991 ; Thompson, 1990) and bodily
shame (Andrews, 1995, 1997 ; Gilbert, 1989 ; Gilbert,
Pehl, & Allen, 1994), as well as genetic and temperamental
factors (Rutter et al., 1997).

The finding that the Rosenberg self-derogation scale

did not appear to be a significant mediator was perhaps
surprising given that negative self-esteem is hypothesised
to play an important role in mediating the link between
early adversity and vulnerability to depression in adult-
hood (Brown & Moran, 1994 ; Harris, Brown, & Bifulco,
1990). However, the findings from this study suggest that
although negative self characteristics may mediate the
link between child abuse and adult depression, attach-
ment style is a more important mediator in the link
between child abuse and adult relationship abilities.
Interestingly, recent work suggests that insecure attach-
ment leads to depressive symptoms in adulthood in-
directly through its impact on low self-esteem and
dysfunctional attitudes (Roberts, Gotlib, & Kassel, 1996).
Alternatively, the finding that self-esteem was not a
significant mediator in the link between child abuse and
adult relationship abilities may in part reflect issues in the
measurement of self-esteem. For example, in an attempt

to overcome the perceived limitations of questionnaire
measures of global self-esteem, Brown and his colleagues
have developed an interview known as the Self-Evalu-
ation and Social Support Instrument (SESS), where
ratings are made on the basis of details about the quality
of the subjects’ life (Brown, Andrews, Bifulco, & Veiel,
1990). They have also found that the SESS is more
effective in predicting depression than Rosenberg’s self-
esteem questionnaire and they suggest that this is because
the SESS taps specific areas of self-dissatisfaction in real-
life situations and is less vulnerable to mood-state effects
than the more global questionnaire measure (Andrews &
Brown, 1993). In order to address these issues, inter-
viewer-based measures of the self such as the SESS could
be incorporated into future studies.

The findings in relation to the attributional measures

also deserve further consideration. Patterns of causal and
responsibility attributions were significantly related to
investigator-based ratings of functioning in the domain
of adult love relationships and the causal and responsi-
bility scales were also found to be moderately related to
child abuse. In particular, women with negative func-
tioning in their love relationships were found to have
more maladaptive causal and responsibility attributions
than women with good functioning in this domain. These
findings add to the growing body of literature on the links
between attributional processes and the quality of in-
timate adult relationships (Baucom, 1987 ; Bradbury &
Fincham, 1990). Although our findings suggest that
attachment processes are more powerful mediators than
relationship attributions in the link between child abuse
and adult relationship abilities, this pattern of findings
may in part be related to a number of methodological
difficulties associated with the measurement of attribu-
tional processes in close relationships. First, a number of
researchers have highlighted the problems associated
with trying to untangle the association between attri-
bution processes and current relationship satisfaction
(Berscheid, 1994 ; Bradbury & Fincham, 1990). Second,
the use of hypothetical situations to assess attributional
processes in close relationships has also been criticised
(Berscheid, 1994). In order to begin to address these
issues future studies may need to incorporate other
techniques that have been developed to assess attributions
in close relationships. These include coding attributions
from videotaped marital interactions (Holtzworth-
Munroe & Jacobsen, 1988) and thought sampling during
current interactions using an intercom (Berman, 1988).

Results from this study also revealed significant associ-

ations between measures of relationship attributions and
attachment style. This suggests that internal working
models of attachment are related to the way in which
adults interpret and attribute meaning to the behaviour
of their partners in close relationships (Bowlby, 1973).
Although the cross-sectional nature of the current study
was not able to address the direction of effects between
these two sets of constructs, the finding that attachment
style, but not attributional processes, were related to
childhood experiences suggests that attachment processes
may be playing a role in influencing the organisation of
attributional processes. However, it seems likely that
once established, internal working models of attachment
and attributional processes will influence one another in a

474

G. M

CARTHY and A. TAYLOR

reciprocal manner, with both sets of cognitive processes
acting to maintain the organisation of the other.

Interpretation of the present findings needs to be

qualified by several factors. First, it is possible that the
procedures used here to assess childhood experiences may
not be an accurate reflection of the quality of these early
experiences. Adult accounts of experiences from child-
hood may be affected by memory bias. However, a recent
review concluded that as long as retrospective reports are
restricted to factual accounts of significant episodes
occurring after the period of infantile amnesia they are
likely to be reasonably valid (Brewin, Andrew, & Gotlib,
1993). These authors suggested that problems associated
with recall bias can be minimised by using investigator-
based interviews. They also found that there was no
convincing evidence that retrospective reports of child-
hood experiences are compromised by current psychiatric
status.

Second, the cross-sectional design of this study makes

it difficult to interpret the direction of causality operating
between attachment style and adult relationship diffi-
culties. Although attachment style may play a role in
determining abilities in close relationships it is also
possible that people who have experienced negative close
relationships might consequently develop a negative
attachment style. In order to address these issues longi-
tudinal prospective studies need to be undertaken where
measures of attachment representations are taken prior
to the assessment of adult relationship abilities.

Third, some circumspection is required in evaluating

this exploratory study given the sample size. Although we
have found significant group differences using standard
levels of significance we have also found smaller group
differences, which might potentially be important even
though they are nonsignificant. This is a consequence of
the limited power of the study. For example, the
difference in BDI levels between women with relationship
difficulties and women with good relationships just failed
to reach nominal significance levels even though the
group differences were quite large. It is therefore im-
portant to replicate the findings of this research in a larger
sample to establish their generality.

If replicated, the finding that attachment processes

may play a role in mediating the link between abusive
childhood experiences and adult relationship difficulties
raises the possibility of designing interventions aimed at
breaking intergenerational cycles of disadvantage. It
suggests that interventions aimed at helping women who
have been abused to become aware of the origins and
potential consequences of their working models of
attachment may be useful in helping them develop
more supportive adult love relationships. Insights from
attachment theory suggest that therapeutic inter-
ventions aimed at helping abused individuals ‘‘ work
through ’’ unresolved childhood experiences could reduce
the likelihood of problems in later intimate relationships.
Alternatively, educational programmes aimed at teaching
at-risk groups about the impact of attachment processes
on intimate relationships and family life may be sufficient
to change the organisation of internal working models of
attachment. More research is required to investigate the
conditions under which representational models of at-
tachment can develop and become updated (Hamilton,

1987). Adolescents who have experienced abusive child-
hood experiences may be a group who would particularly
benefit from such interventions. During adolescence
young people are beginning to establish romantic rela-
tionships and they have also acquired the necessary
cognitive skills to enable them think about the causes and
consequences of their own and others’ behaviour.

Acknowledgements

—We would like to thank the women who

generously agreed to take part in this study, Dr Barbara
Maughan for her helpful comments on earlier drafts of this
paper, and Denise Shields for her help in preparing the
manuscript.

References

Ainsworth, M., Blehar, M., Waters, E., & Wall, S. (1978).

Patterns of attachment : A psychological study of the Strange
Situation

. Hilldale, NJ : Lawrence Erlbaum.

Alexander, P. (1993). The differential effects of abuse charac-

teristics and attachment in the prediction of long-term effects
of sexual abuse. Journal of Interpersonal Violence, 8, 346–362.

Andrews, B. (1992). Attribution process in victims of marital

violence : Who do women blame and why ? In J. H. Harvey,
T. L. Orbuch, & A. L. Weber (Eds.), Attributions, accounts
and close relationships

(pp. 176–193). New York : Springer-

Verlag.

Andrews, B. (1995). Bodily shame as a mediator between

abusive experiences and depression. Journal of Abnormal
Psychology

, 104, 277–285.

Andrews, B. (1997). Bodily shame in relation to abuse in

childhood and bulimia : A preliminary investigation. British
Journal of Clinical Psychology

, 36, 41–50.

Andrews, B., & Brewin, C. R. (1990). Attributions for marital

violence : A study of antecedents and consequences. Journal
of Marriage and the Family

, 52, 757–767.

Andrews, B., & Brown, G. W. (1988). Marital violence in the

community : A biographical approach. British Journal of
Psychiatry

, 153, 305–312.

Andrews, B., & Brown, G. W. (1993). Self-esteem and vul-

nerability to depression : The concurrent validity of interview
and questionnaire measures. Journal of Abnormal Psycho-
logy

, 102, 565–572.

Arend, R., Gove, F., & Sroufe, L. A. (1979). Continuity of

individual adaptation from infancy to kindergarten : A
predictive study of ego-resiliency and curiosity in pre-
schoolers. Child Development, 50, 950–959.

Baron, R. M., & Kenny, D. A. (1986). The moderator mediator

variable in social-psychological research : Conceptual, stra-
tegic, and statistical considerations. Journal of Personality
and Social Psychology

, 51, 1173–1182.

Bartholomew, K. (1990). Avoidance of intimacy : An attach-

ment perspective. Journal of Social and Personal Relation-
ships

, 7, 147–178.

Bartholomew, K., & Horowitz, L. M. (1991). Attachment styles

among young adults : A test of a four-category model. Journal
of Personality and Social Psychology

, 61, 226–244.

Baucom, D. H. (1987). Attributions in distressed relations :

How can we explain them ? In S. Duck & D. Perlman (Eds.),
Heterosexual relations

, marriage, and divorce (pp. 177–206).

London : Sage.

Baucom, D. H., Epstein, N., Sayers, S., & Goldman-Sher, T.

(1989). The role of cognitions in marital relationships :
Definitional methodological, and conceptual issues. Journal
of Consulting and Clinical Psychology

, 57, 31–38.

Beck, A. (1984). The Beck Depression Inventory. In M.

Williams (Ed.), The psychological treatment of depression : A

475

AVOIDANT

\AMBIVALENT ATTACHMENT STYLE

guide to the theory and practice of cognitive behavioral therapy

.

New York : Free Press.

Belsky, J., & Cassidy, J. (1994). Attachment theory and

evidence. In M. Rutter & D. F. Hay (Eds.), Developmental
principles and clinical issues in psychology and psychiatry

.

Oxford : Blackwell.

Beitchman, J. H., Zucker, K., & Hood, J. E. (1992). A review of

the long-term consequences of child sexual abuse. Child
Abuse and Neglect

, 16, 101–118.

Berman, W. (1988). The role of attachment in the post-divorce

experience. Journal of Personality, Society and Psychology,
54

, 496–504.

Berscheid, E. (1994). Interpersonal relationships. Annual Review

of Psychology

, 45, 79–129.

Bifulco, A., Brown, G. W., & Adler, Z. (1991). Early sexual

abuse and clinical depression in adult life. British Journal of
Psychiatry

, 159, 115–122.

Bifulco, A., Brown, G. W., & Harris, T. O. (1994). Childhood

Experience of Care and Abuse (CECA) : A retrospective
interview measure. Journal of Child Psychology and Psy-
chiatry

, 35, 1419–1435.

Birtchnell, J. (1993). Does recollection of exposure to poor

maternal care in childhood affect later ability to relate ?,
British Journal of Psychiatry

, 162, 335–344.

Bowlby, J. (1969). Attachment and loss, Vol 1 : Attachment.

London : Penguin.

Bowlby, J. (1973). Attachment and loss, Vol 2 : Separation.

London : Penguin.

Bowlby, J. (1979). The making and breaking of affectional bonds.

London : Tavistock Publications.

Bowlby, J. (1980). Attachment and loss, Vol 3 : Loss, sadness and

depression

. New York : Basic Books.

Bowlby, J. (1988). A secure base : Parent–child attachment and

healthy human development

. New York : Basic Books.

Bradbury, T. N., Beach, S. R. H., Fincham, F. D., & Nelson,

G. M. (1996). Attributions and behavior in functional and
dysfunctional marriages. Journal of Consulting and Clinical
Psychology

, 64, 569–576.

Bradbury, T. N., & Fincham, F. D. (1990). Attributions in

marriage : Review and critique. Psychological Bulletin, 107,
3–33.

Bradbury, T. N., & Fincham, F. D. (1992). Attributions and

behavior in marital interaction. Journal of Personality and
Social Psychology

, 63, 613–628.

Brennan, K. A., Shaver, P., & Tobey, A. (1991). Attachment

styles, gender and parental problem drinking. Journal of
Social and Personal Relationships

, 8, 451–466.

Brewin, C. R., Andrews, B., & Gotlib, I. H. (1993). Psycho-

pathology and early experiences : A reappraisal of retro-
spective reports. Psychological Bulletin, 113, 82–98.

Brown, G. W., Andrews, B., Bifulco, A., & Veiel, H. O. F.

(1990). Self-esteem and depression. I. Measurement issues
and prediction of onset. Social Psychiatry and Psychiatric
Epidemiology

, 25, 200–209.

Brown, G. W., Bifulco, A., & Andrews, B. (1990). Self-esteem

and depression. III. Aetiological issues. Social Psychiatry and
Psychiatric Epidemiology

, 25, 235–243.

Brown, G. W., Bifulco, A., Veiel, H. O. F., & Andrews, B.

(1990). Self-esteem and depression. II. Social correlates of
self-esteem. Social Psychiatry and Psychiatric Epidemiology,
25

, 225–234.

Brown, G. W., & Harris, T. O. (1978). Social origins of

depression

. London : Tavistock.

Brown, G. W., & Moran, P. (1994). Clinical and psychosocial

origins of chronic depressive episodes I : A community survey.
British Journal of Psychiatry

, 165, 447–456.

Brown, G. W., & Rutter, M. L. (1966). The measurement of

family activities and relationships : A methodological study.
Human Relations

, 19, 241–263.

Carlson, V., Cicchetti, D., Barnett, D., & Braunwald, K. (1989).

Disorganized

\disoriented attachment relationships in mal-

treated infants. Developmental Psychology, 25, 525–531.

Champion, L., Goodall, G., & Rutter, M. (1995). Behaviour

problems in childhood and stressors in early adult life : I. A
20-year follow-up of London school children. Psychological
Medicine

, 25, 231–246.

Cicchetti, D., Ganiban, J., & Barnett, D. (1991). Contributions

from the study of high risk populations to understanding the
development of emotion regulation. In K. Dodge & J. Garber
(Eds.), The development of emotion regulation (pp. 15–48).
New York : Cambridge University Press.

Cicchetti, D., & Toth, S. L. (1995). Developmental psycho-

pathology and disorders of affect. In D. Cicchetti & S. L.
Toth (Eds.) Developmental psychopathology, Vol. 2 : Risk,
disorder, adaption.

New York : Wiley.

Collins, N. L., & Read, S. J. (1990). Adult attachment, working

models, and relationship quality in dating couples. Journal of
Personality and Social Psychology

, 58, 644–663.

Collins, N. L., & Read, S. J. (1994). Cognitive representations

of attachment : The structure and function of working
models. In K. Bartholomew & D. Perlman (Eds.), Advances
in personal relationships

, Vol. 5 : Attachment processes in

adulthood

(pp. 53–90). London : Jessica Kingsley.

Crittenden, P. M. (1988). Relationships at risk. In J. Belsky &

T. Nezworski (Eds.), Clinical implications of attachment.
Hillsdale, NJ : Lawrence Erlbaum.

Crowell, J. A., & Treboux, D. (1995). A review of adult

attachment measures : Implications for theory and research.
Social Development

, 4, 3.

Davis, K., Kirkpatrick, L., Levy, M., & O’Hearn, R. (1994).

Stalking the elusive love style : Attachment styles, love styles
and relationship developments. In R. Erber & R. Gilmour
(Eds.), Theoretical frameworks for personal relationships.
Hillsdale, NJ : Lawrence Erlbaum.

Dodge, K. A., Bates, J. E., & Pettit, G. S. (1990). Mechanisms

in the cycle of violence. Science, 250, 1678–1683.

Dodge, K. A., Pettit, G. S., Bates, J. E., & Valente, E. (1995).

Social-information-processing patterns mediate the effect of
early physical abuse on later conduct problems. Journal of
Abnormal Psychology

, 104, 632–643.

Egeland, B., & Sroufe, L. A. (1981). Developmental sequelae

of maltreatment in infancy. In R. Rizley & D. Cicchetti
(Eds.), Developmental perspectives in child maltreatment. San
Francisco, CA : Jossey-Bass.

Emery, R. E. (1982). Interparental conflict and the children of

discord and divorce. Psychological Bulletin, 92, 310–330.

Erikson, E. (1963). Childhood and society. New York : Norton.
Farrington, D. F. (1995). The Twelfth Jack Tizard Memorial

Lecture. The development of offending and antisocial be-
haviour from childhood : Key findings from the Cambridge
Study in delinquent development. Journal of Child Psy-
chology and Psychiatry

, 36, 929–964.

Feeney, J., & Noller, P. (1990). Attachment style as a predictor

of adult romantic relationships. Journal of Personality and
Social Psychology

, 8, 281–291.

Fergusson, D. M., & Horwood, L. J. (1995). Predictive validity

of categorically and dimensionally scored measures of dis-
ruptive behaviour disorders. American Academy of Child and
Adolescent Psychiatry

, 34, 477–487.

Fergusson, D. M., Horwood, L. J., & Lynskey, M. T. (1996).

Childhood sexual abuse and psychiatric disorder in young
adulthood : II. Psychiatric outcomes of childhood sexual
abuse. Journal of the American Academy of Child and
Adolescent Psychiatry

, 34, 1365–1374.

Fergusson, D. M., Lynskey, M. T., & Horwood, L. J. (1996).

Childhood sexual abuse and psychiatric disorder in young
adulthood : I. Prevalence of sexual abuse and factors associ-

476

G. M

CARTHY and A. TAYLOR

ated with sexual abuse. Journal of the American Academy of
Child and Adolescent Psychiatry

, 34, 1355–1364.

Fincham, F. D., Beach, S. R., & Nelson, G. (1987). Attribu-

tional processes in distressed and nondistressed couples : 3.
Causal and responsibility attributions for spouse behavior.
Cognitive Therapy and Research

, 11, 71–86.

Fincham, F. D., & Bradbury, T. N. (1992). Assessing attribu-

tions in marriage : The relationship attribution measure.
Journal of Personality and Social Psychology

, 62, 457–468.

Finkelhor, D. (1983). Common features of family abuse. In D.

Finkelhor, R. Gelles, G. Hotaling, & M. Straus (Eds.), The
dark side of families : Current family violence research

.

London : Sage.

Finkelor, D., Hotaling, G., Lewis, I. A., & Smith, C. (1990).

Sexual abuse in a national survey of adult men and women :
Prevalence, characteristics, and risk factors. Child Abuse and
Neglect

, 14, 19–28.

Fromouth, M. E. (1986). The relationship of childhood sexual

abuse with later psychological and sexual adjustment in a
sample of college women. Child Abuse and Neglect, 10, 5–15.

Gilbert, P. (1989). Human nature and suffering. London :

Lawrence Erlbaum.

Gilbert, P., Pehl, J., & Allan, S. (1994). The phenomenology of

shame and guilt : An empirical investigation. British Journal
of Medical Psychology

, 67, 23–36.

Gold, E. R. (1986). Long-term effects of sexual victimization in

childhood : An attributional approach. Journal of Consulting
and Clinical Psychology

, 54, 471–475.

Hamilton, V. (1987). Some problems in the clinical application

of attachment theory. Psychoanalytic Psychotherapy, 3,
67–83.

Harris, T., Brown, G. W., & Bifulco, A. (1990). Loss of a parent

in childhood and adult psychiatric disorder : A tentative
overall model. Developmental Psychopathology, 2, 311–328.

Harter, S., Alexander, P. C., & Neimeyer, R. A. (1988). Long-

term effects of incestuous child abuse in college women :
Social adjustment, social cognition and family characteristics.
Journal of Consulting and Clinical Psychology

, 56, 5–8.

Hazan, C., & Shaver, P. (1987). Romantic love conceptualized

as an attachment process. Journal of Personality and Social
Psychology

, 52, 511–524.

Hill, J., Harrington, R., Fudge, H., Rutter, M., & Pickles, A.

(1989). Adult Personality Functioning Assessment (APFA) :
An investigator-based standardized interview. British Journal
of Psychiatry

, 155, 24–35.

Holtzworth-Munroe, A. (1992). Social skill deficits in maritally

violent men : Interpreting the data using a social information
processing model. Clinical Psychology Review, 12, 605–617.

Holtzworth-Munroe, A., & Anglin, K. (1991). The competency

of responses given by maritally violent versus nonviolent men
to problematic marital situations. Violence and Victims, 6,
257–269.

Holtzworth-Munroe, A., & Hutchinson, G. (1993). Attributing

negative intent to wife behavior : The attributions of maritally
violent versus nonviolent men. Journal of Abnormal Psy-
chology

, 102, 206–211.

Holtzworth-Munroe, A., & Jacobson, N. S. (1988). Toward a

methodology for coding spontaneous causal attributions :
Preliminary results with married couples. Journal of Social
and Clinical Psychology

, 7, 101–112.

Jouriles, E. N., Piffner, L. J., & O’Leary, S. G. (1988). Marital

conflict, parenting, and toddler conduct problems. Journal of
Abnormal Child Psychology

, 16, 197–206.

Kaplan, H. B., & Pokorny, A. D. (1969). Self-derogation and

psycho-social adjustment. Journal of Nervous and Mental
Disease

, 149, 421–434.

Kohn, M. L., & Schooler, C. (1983). Work and personality : An

inquiry into the impact of social stratification.

Norwood, NJ :

Ablex.

Kohut, H. (1977). The restoration of the self. New York :

International Universities Press.

Latty-Mann, H. (1989). The influence of family of origin

dynamics on the etiology of adult attachment styles and love
styles : With special focus on adult children of alcoholics.
Unpublished Master’s thesis, University of South Carolina,
Columbia, SC.

Latty-Mann, H. (1991). An intergenerational approach to

studying the influence of family of origin dynamics on the
etiology of adult attachment styles : With special focus on adult
children of alcoholics.

Unpublished doctoral dissertation,

University of South Carolina, Columbia, SC.

Latty-Mann, H., & Davis, K. E. (1988). Adult children of

alcoholics : Family dynamics and patterns of romantic relating

.

Presented at the International Conference on Personal
Relationships, Vancouver, Canada.

Latty-Mann, H., & Davis, K. E. (1996). Attachment theory and

partner choice : Preference and actuality. Journal of Social
and Personal Relationships

, 13, 5–23.

Lewinsohn, P. M., & Rosenbaum, M. (1987). Recall of parental

behaviour by acute depressives, remitted depressives, and
nondepressives. Journal of Personality and Social Psychology,
52

, 611–619.

LogXact. (1996). LogXact for Windows : User manual. Cam-

bridge, MA : Cytel Software Corporation.

Lynch, M., & Cicchetti, D. (1991). Patterns of relatedness in

maltreated and nonmaltreated children : Among multiple
representational models. Development and Psychopathology,
3

, 207–226.

Lyons-Ruth, K., Alpern, L., & Repacholi, B. (1993). Dis-

organized infant attachment classification and maternal
psychosocial problems as predictors of hostile-aggressive
behaviour in the classroom. Child Development, 64, 572–585.

Lyons-Ruth, K., Connell, D. B., Grunebaum, H., & Botein, S.

(1990). Infants at social risk : Maternal depression and family
support services as mediators of infant development and
security of attachment. Child Development, 61, 85–98.

Main, M. (1991). Metacognitive knowledge, metacognitive

monitoring and singular (coherent) versus multiple (inco-
herent) models of attachment : Findings and directions for
future research. In C. M. Parkes, J. Stevenson-Hinde, & P.
Marris (Eds.), Attachment across the life cycle (pp. 127–159)
London : Routledge.

Main, M., Kaplan, N., & Cassidy, J. (1985). Security in infancy,

childhood and adulthood : A move to the level of representa-
tion. In I. Bretherton & E. Waters (Eds.) Growing points of
attachment theory and research

. Monographs of the Society for

Research in Child Development

, 50, 66–104.

Main, M., & Soloman, J. (1986). Discovery of an insecure-

disorganized

\disorientated attachment pattern: Procedures,

findings and implications for the classification of behaviour.
In T. B. Brazelton & M. Yogman (Eds.), Affective devel-
opment in infancy

. Norwood, NJ : Ablex.

Malinosky-Rummell, R., & Hansen, D. J. (1993). Long-term

consequences of childhood physical abuse. Psychological
Bulletin

, 114, 68–79.

Matas, L., Arend, R., & Sroufe, L. A. (1978). Continuity of

adaptation in the second year. The relationship between
quality of attachment and later competent functioning. Child
Development

, 49, 547–556.

Maughan, B., & Hagell, A. (1996). Poor readers in adulthood :

Psychosocial functioning. Development and Psychopathology,
8

, 457–476.

Maughan, B, & McCarthy, G. (1997). Childhood adversities

and psychosocial disorders. British Medical Bulletin, 53,
156–169.

Mehta, C. R., & Patel, N. R. (1995). Exact logistic regression :

Theory and examples. Statistics in Medicine, 14, 2143–2160.

477

AVOIDANT

\AMBIVALENT ATTACHMENT STYLE

Mikulincer, M., & Erev, I. (1991). Attachment style and the

structure of romantic love. British Journal of Social Psy-
chology

, 30, 273–291.

Mikulincer, M., & Nachshon, O. (1991). Attachment styles and

patterns of self-disclosure. Journal of Personality and Social
Psychology

, 61, 321–331.

Mullen, P. E., Martin, J. L., Anderson, J. C., Romans, S. E., &

Herbison, G. P. (1994). The effect of child sexual abuse on
social, interpersonal and sexual function in adult life. British
Journal of Psychiatry

, 165, 35–47.

Nurmi, J. E., & Pulliainen, H. (1991). The changing parent–

child relationship, self-esteem, and intelligence as determi-
nants of orientation to the future during early adolescence.
Journal of Adolescence

, 14, 35–51.

Paradise, J. E., Rose, L., Sleeper, L. A., & Nathanson, M.

(1994). Behaviour, family functioning, school performance
and predictors of persistent disturbance in sexually abused
children. Pediatrics, 93, 452–459.

Patterson, G. R. (1982). A social learning approach, Vol. 3.

Coercive family process.

Eugene, OR : Castalia Publishing

Co.

Quinton, D., Pickles, A., Maughan, B., & Rutter, M. (1993).

Partners, peers and pathways : Assortative pairing and
continuities in conduct disorder. Development and Psycho-
pathology

, 5, 763–783.

Quinton, D., & Rutter, M. (1988). Parental breakdown : The

making and breaking of intergenerational links

. Aldershot,

U.K. : Gower.

Roberts, J. E., Gotlib, I. H., & Kassel, J. D. (1996). Adult

attachment security and symptoms of depression : The
mediating roles of dysfunctional attitudes and low self-
esteem. Journal of Personality and Social Psychiatry, 70,
310–320.

Rosenburg, D. (1984). The quality and content of preschool

fantasy play : Correlates in concurrent social

\personality

function and early mother-child attachment relationships

.

Unpublished doctoral dissertation, University of Minnesota.

Rosenberg, M. (1965). Society and the adolescent self-image.

Princeton, NH : Princeton University Press.

Rutter, M. (1989). Pathways from childhood to adult life.

Journal of Child Psychology and Psychiatry

, 30, 23–51.

Rutter, M., Dunn, J., Plomin, R., Simonoff, E., Pickles, A.,

Maughan, B., Ormel, J., Meyer, J., & Eaves, L. (1997).
Integrating nature and nurture : Implications for person–
environment correlations and interactions for developmental
psychopathology. Development and Psychopathology, 9, 335–
364.

Sampson, R., & Laub, J. (1993). Crime in the making : Pathways

and turning points through life

. Cambridge, MA : Harvard

University Press.

Schneider-Rosen, K., Braunwald, K. G., Carlson, V., & Cic-

chetti, D. (1985). Current perspectives in attachment theory :
Illustration from the study of maltreated infants. In I.
Bretherton & E. Waters (Eds.), Growing points in attachment
theory and research (pp. 194–210). Monographs of the Society
for Research and Development in Child Development

, 50 (1–2,

Serial No. 209).

Schneider-Rosen, K., & Cicchetti, D. (1984). The relationship

between affect and cognition in maltreated infants : Quality of
attachment and the development of visual self-recognition.
Child Development

, 55, 648–658.

Shaver, P. R., & Hazan, C. (1993). Adult romantic attachment :

Theory and evidence. Advances in Personal Relationships, 4,
29–70.

Simpson, J. A. (1990). Influence of attachment styles on

romantic relationships. Journal of Personality and Social
Psychology

, 59, 971–980.

Simpson, J. A., Rholes, W. S., & Nelligan, J. S. (1992). Support-

seeking and support-giving within couple members in an
anxiety-provoking situation : The role of attachment styles.
Journal of Personality and Social Psychology

, 62, 434–446.

SPSS Inc. (1994). SPSS for Windows V. 6.1. Chicago, IL :

Author.

Sroufe, L. A. (1983). Infant–caregiver attachment and adap-

tation in the preschool : The roots of competence and
maladaptation. In M. Pearlmutter (Ed.), Minnesota Sym-
posium in Child Psychology

, Vol. 16. Hillsdale, NJ : Lawrence

Erlbaum.

Sroufe, L. A. (1986). Appraisal : Bowlby’s contribution to

psychoanalytic theory and developmental psychology ; at-
tachment : separation : loss. Journal of Child Psychology and
Psychiatry

, 27, 841–849.

Sroufe, L. A. (1989). Pathways to adaptation and maladap-

tion : Psychopathology as developmental deviation. In D.
Cicchetti (Ed.) Rochester Symposium on Developmental Psy-
chopathology

, Vol. 1 : The emergence of a discipline (pp.

13–40). Hillsdale, NJ : Lawrence Erlbaum.

Sullivan, H. S. (1953). The interpersonal theory of psychiatry.

New York : Norton.

Thompson, R. (1990). Emotions and self-regulation. Nebraska

Symposium on Motivation

, 36, 367–467.

Urban, J., Carlson, E., Egeland, B., & Sroufe, L. A. (1991).

Patterns of individual adaptation across childhood. Develop-
ment and Psychopathology

, 4, 445–460.

Vondra, J., Barnett, D., & Cicchetti, D. (1989). Perceived and

actual competence among maltreated and comparison school
children. Development and Psychopathology, 1, 237–255.

Waters, E., Wippman, J., & Sroufe, L. A. (1979). Attachment,

positive affect, and competence in the peer group : Two
studies in construct validation. Child Development, 50,
821–829.

Weiss, B., Dodge, K. A., Bates, J. E., & Pettit, G. S. (1992).

Some consequences of early harsh discipline : Child ag-
gression and a maladaptive social information processing
style. Child Development, 63, 1321–1335.

Wyatt, G. E., & Mickey, M. R. (1987). Ameliorating the effects

of child sexual abuse : An exploratory study of support by
parents and others. Journal of Interpersonal Violence, 2,
403–414.

Zeller, R. A., & Carmaines, E. G. (1980). Measurement in the

social sciences : The link between theory and data.

Cambridge :

Cambridge University Press.

Manuscript accepted 30 June 1998

This document is a scanned copy of a printed document. No warranty is given about the accuracy of the copy.

Users should refer to the original published version of the material.