Sie sind auf Seite 1von 15

Journal of Child Psychology and Psychiatry 46:2 (2005), pp 115–127 doi: 10.1111/j.1469-7610.2004.00396.

Practitioner Review: The contribution of


attachment theory to child custody assessments
James G. Byrne,1 Thomas G. O’Connor,2 Robert S. Marvin,3
and William F. Whelan3
1
Lucena Clinic, Dublin, Ireland; 2University of Rochester, Rochester, USA;
3
University of Virginia, Charlottesville, VA, USA

Background: The area of child custody assessments continues to fail to meet the evidence-based
threshold now established in clinical practice. This is despite the existence, for many years, of published
guidelines governing the practice of custody assessments available from a number of professional
bodies. Methods: This article reviews the potential of attachment theory to contribute to the concep-
tualization of custody evaluations, clinical assessment, and the development of evidence-based
practice. Particular attention is paid to specific instruments used to assess attachment in clinic and
non-clinic settings. Results: Guidelines concerning child custody assessments highlight the particular
importance of assessing attachment and parent–child relationship quality. However, measures often
used in the course of a custody assessment are not backed up with empirical research, and the
measures that are supported by empirical research have been slow to influence practice. There may be
conceptual and measurement advantages of considering an attachment research-informed custody
assessment. Discussion: Attachment theory has obvious conceptual relevance for the child custody
context. Further clinical research is needed to demonstrate the usefulness of attachment research
measures; research of this kind may shed important light on the development and resilience of affec-
tional bonds. Keywords: Attachment, custody evaluation, assessment.

Evidence-based practice has come to be accepted as determine the custody decision, particularly the role
a standard to be achieved in the practice of psychi- of psychological input, 3) consider the potential of
atry and psychology. Application to child and ado- attachment theory for informing practice by review-
lescent mental health services is no exception, ing clinical and research assessment strategies and
although it is recognized that progress in this area is evaluating how attachment theory has been integ-
just beginning (Ramchandani, Joughlin, & Zwi, rated so far, and 4) provide practical recommenda-
2001). Progress in achieving the evidence-based tions for developing an attachment-oriented
standard in forensic child and adolescent mental evidence-based model of custody evaluation.
health has been slower still. In particular, despite
many years of clinical experience and despite the
existence of practice guidelines from many profes- 1. Guidelines concerning current practice
sional bodies, most clinicians would readily accept
Estimates of the number or percentage of parents
that there is a paucity of relevant research evidence on
with children seeking divorce that require a court
which to base the practice of child custody evalu-
hearing and a psychological evaluation to inform
ations.1 This is all the more relevant given that the
custody determination are difficult to obtain. Sim-
courts are moving from a sociologically based view of
ilarly, for the increasing numbers of parents seeking
truth (‘Truth is what experts in the field believe’) to-
alternative dispute resolution, such as mediation,
wards a method-based view (‘Truth is what is dis-
the degree to which outside psychological input is
covered through the appropriate application of
needed is difficult to discern. Whatever the figure
scientific methods’; Galatzer-Levy & Ostrov, 1999).
may be in a particular country or jurisdiction, it is
The aims of this article are to 1) review briefly
likely to be small because only a fraction of cases
guidelines concerning custody (or residence) deter-
include outside psychological evaluation. However, it
mination and the role that attachment theory has
is almost assured that those cases that do proceed
come to play, 2) discuss the factors that may
through the court system and require a psycholo-
gical evaluation are the more difficult and compli-
1
The focus in this article is on custody evaluation in the con- cated cases, characterized by contested views of
text of separation/divorce rather than child protection in the custody, current and past conflict between separ-
case of maltreatment or other family problems, although some
ating partners, and inability to negotiate emotionally
of the ideas and implications are similar in both kinds of
custody determination. In addition, we refer in this article to
challenging and upsetting matters; accusations by
‘custody’ because that is the term most often used interna- one or both parents of mistreatment of the child are
tionally, but other terms (e.g., ‘residence’ in the UK) are also also not uncommon (Halikias, 1994; Johnston &
used. Campbell, 1993). Parents who are working together
Ó Association for Child Psychology and Psychiatry, 2004.
Published by Blackwell Publishing, 9600 Garsington Road, Oxford OX4 2DQ, UK and 350 Main Street, Malden, MA 02148, USA
116 James G. Byrne et al.

for the interests of the child(ren) are unlikely to re- that it is now common in practice to refer to attach-
quire the input of the court system and a psycholo- ment, but how attachment is assessed varies widely
gical evaluation (although sometimes consultation and, in some cases, is contrary to what is known
with a mental health professional is sought). from clinical and developmental research. Below, we
The principal objective of the custody assessment discuss some of the difficulties connected with the
is to provide valid information to the court on the popularization of attachment theory in child custody
current and future impact on the child and family of evaluation and then discuss how attachment theory
alternative custody options. Practice parameters for might be successfully incorporated into current
meeting this objective have been published by nu- practice. However, we first outline the broader con-
merous professional organizations2 and custody text for understanding how psychological theory may
evaluation practices of various mental health pro- affect custody determination.
fessionals have been reviewed (Ackerman & Acker-
man, 1997; Jameson, Ehrenberg, & Hunter, 1997;
Keilin & Bloom, 1986). Practice parameters reflect
2. Standards, theories, and research findings
substantial agreement about what, in general,
and their influence on custody determination
should be assessed in a custody evaluation. Ex-
amples include parental mental health, parenting Determination of custody is ultimately a legal
style or parent–child interactions, the psychological decision. A much-debated topic is how much scope
and developmental needs of the child, and the ‘fit’ there is or should be for psychological research and
between these latter two. Practice parameters also theory into the decision-making process, and how
typically recommend clinical interviews with each prepared psychologists/psychiatrists are to con-
parent and with the child, as well as an observation tribute to this process (Kunin, Ebbesen, & Konecni,
of the interaction between each adult and child. 1992; Melton, Petrika, Poythress, & Slobogin, 1987).
Furthermore, existing practice guidelines also agree Custody determination will invariably reflect legal
that conclusions and recommendations should be practice and statutes, which in turn mirror political
based, inter alia, on scientific procedures. Yet, and social expectations. This is underscored by an
practice parameters stop short of describing what historical view of custody determination which re-
these scientific procedures are or should be. veals that substantial changes in legal practice occur
Although most guidelines avoid the promotion of in the absence of supporting evidence (see below).
particular theoretical perspectives, child-parent Furthermore, how the different inputs into the cus-
attachment is singled out by some (e.g., ‘The tody decision are weighed is hard to estimate, may be
assessment of the quality of the attachments be- idiosyncratic, and will almost certainly vary across
tween the parents and the children is the centerpiece jurisdiction. Indeed, this variation is amply demon-
of the evaluation’; AACAP, 1997). Nevertheless, how strated among the different settings in which the
attachment is assessed is not discussed and what is authors work. The implication is that, to an im-
meant by ‘attachment’ may be uncertain. Thus, for portant degree, a range of considerations weight le-
example, the American Academy of Child and Ado- gal decisions aside from whatever lessons
lescent Psychiatry guidelines recommend that the psychological research may provide. Accordingly, it
evaluator should consider the child’s ‘level of may be unreasonable to expect that research find-
attachment with adult figures.’ The difficulty here is ings’ impact on child custody decisions would be
that the notion of a ‘level’ of attachment is not readily either immediate or direct.
interpretable from existing theory or research (in
contrast to, say, ‘Secure’ or ‘Insecure’ patterns of Standards. A number of legal standards – many of
child–parent attachments). The more general point is which have been incorporated into statute – have
been applied to child custody determinations. These
2
Child custody evaluation guidelines are available from: include a) the ‘tender years’ presumption that
American Psychiatric Association. (1982). Child custody con- younger children would be best placed with their
sultation: Report of the task force on clinical assessment in mother, b) the ‘best interests of the child’ standard
child custody. Washington, DC: American Psychiatric Associ-
which moved decision-making toward child-focused
ation; American Psychological Association. (1994). Guidelines
for child custody evaluations in divorce proceedings. American
rather than parent-focused concerns, c) the notion of
Psychologist, 49, 677–680; the ‘least detrimental alternative’ which emphasized
American Academy of Child and Adolescent Psychiatry. (1997). the need to balance risks and benefits of alternative
Practice parameters for child custody evaluation. Journal of the options, d) the ‘primary caretaker’ presumption and,
American Academy of Child and Adolescent Psychiatry, 36, more recently, e) a tendency to favor joint over sole
57s–68s; custody, and f) an increased tendency to favor equal
Association of Family and Conciliation Courts (AFCC). (1994).
physical custody (Bartlett, 1999; Derdeyn & Scott,
Model standards of practice for child custody evaluation.
Family Conciliation Courts, 32, 504–513;
1984; Goldstein, Freud, & Solnit, 1980; Kelly, 1994;
Group for the Advancement of Psychiatry [GAP], Committee on Maccoby & Mnookin, 1992). These standards have
the Family. (1981). Divorce, child custody and the family. San been criticized in the psychological and psychiatric
Francisco: Jossey-Bass. literatures for being vague. That may be partly
Practitioner Review: The contribution of attachment theory to child custody assessments 117

explained by the fact that, although these standards not be valid when used in a custody evaluation. This
are often couched in psychological language, they is another way of saying that there are dangers of
are instead legal terms with no particular evidence using psychological tests for purposes for which they
base or theoretical background. Indeed, how, or if, were not developed, as noted in the American Psy-
these changing standards are reflected in the prac- chological Association’s (1994) guidelines. Addition-
tice of child custody evaluations is uncertain. ally, some of the more commonly used measures for
Therefore, rather than attach psychological signific- children have questionable validity; one such
ance to these terms, it may be best to view these example is projective tests (see, Klein, 1986). A fur-
terms as default positions for custody determina- ther limitation is that some measures routinely used
tions to be relied on in the absence of overwhelming in custody evaluations may be valid (e.g., IQ tests),
contrary evidence, such as that one parent is unfit. but may not serve an obvious function in most cases.
Another conclusion from survey studies is that
Psychological theory. Theories of child development there is a surprising lack of cross-fertilization from
do not figure centrally in discussions of custody other lines of clinical-developmental research. One
assessments. Instead, psychological input into child area that stands out is the parent–child relationship.
custody determinations is often based on practical The implication, from practice surveys, that evid-
considerations (e.g., work schedules of parents) and ence-based measures for assessing parent–child
generic mental health and relationship assessments relationship have no discernible place in the current
(e.g., evidence of parental mental illness or of a his- practice of child custody evaluations is striking given
tory of abusive/neglectful parental care). The the availability of valid and reliable measures of
unspecified conceptual basis underlying psycholo- parent–child relationships and attachment (O’Con-
gical assessments is inevitable given the absence of nor, 2002). On the other hand, measures of the
relevant research. One consequence of this is that, if parent–child relationship that have been developed
there were contrasting views expressed by experts of for, and are used in, custody evaluations (Ackerman
opposing parties, there would be no sound basis for & Schoendorf, 1992; Bricklin, 1990; Gerard, 1994;
favoring one approach over another. Subsequent McDermott, Tseng, Char, & Fukunaga, 1978) have
sections of this article discuss the potential role of not been adequately scrutinized in research. Fur-
attachment theory as a conceptual framework for thermore, some measures developed for use in cus-
assessment and decision-making. tody evaluation do not include methods for observing
parent–child interactions, despite the fact that
Research findings relevant to custody evalua- practice parameters recommend that this is an
tions. Psychological evaluation and court coun- important part of the evaluation.
selors may play a decisive role in child custody Thus, the current situation is that measures of the
determination (see, Kunin et al., 1992), but the parent–child relationship with extensive validation
empirical base on which these clinical assessments are not generally used in clinical settings, and that
are carried out is the subject of remarkably little measures developed for custody evaluations have
systematic research. Studies that do exist are con- not been subjected to empirical scrutiny. One sug-
cerned with surveying measures used by practi- gestion for bridging clinical need and research
tioners. These findings demonstrate what common expertise in custody evaluations is the inclusion of
practice is rather than how reliable and valid these an attachment research model. The remainder of
measures are. One conclusion from survey studies is this article focuses on the prospects and challenges
that there is widespread use of personality, psycho- of such an approach.
pathology, and intelligence assessments (Ackerman
& Ackerman, 1997; Heinze & Grisso, 1996; Hysju-
lien, Wood, & Benjamin, 1994; Keilin & Bloom,
3. Attachment theory and its relevance to
1986; O’Donohue & Bradley, 1999; Quinnell & Bow,
custody assessments
2001). This kind of information is valuable for
making judgments about the mental health and As noted, the term ‘attachment’ is in regular use in
psychological resources of parents and children, but the custody context and is explicitly referred to in
does not provide sensitive information on issues some guidelines. For example, in his recent review,
more fundamental to the custody evaluation, notably Bernet (2002) explicitly states that the child’s
parenting competence or parent–child relationship attachment to each parent should be assessed and
quality. An additional concern is that measures of that this information should feature in the written
parental personality or child behavioral problems report. However, our reading of the available litera-
that have been validated in population and clinical ture and experience with clinical and legal colleagues
samples may be subject to particularly strong par- has taught us that, in many cases, the use of the
ental reporter bias when used in child custody term ‘attachment’ is not meant to convey the theory
evaluation (Hysjulien et al., 1994). The general point and research tradition of Bowlby and Ainsworth; in
is that measures that have been shown to be reliable other cases there is a clear misunderstanding and
and valid outside a custody evaluation context may misapplication of attachment theory ideas. The
118 James G. Byrne et al.

current state of confusion means that before we Secure attachment relationship provides the child
discuss how an attachment theory perspective may with resilience to cope with stress; in contrast, an
inform child custody evaluations, we need first to Insecure attachment places the child at risk for
clarify what we mean by attachment and then specify subsequent developmental problems (e.g., problems
evidence-based assessment measures. in peer relationships in middle childhood) via a
Attachment theory, developed by Bowlby (1973, number of hypothesized mechanisms.
1980, 1982), and buttressed by considerable evid- Insecure attachment is not synonymous with
ence from several decades of research of Ainsworth pathology, just as a Secure attachment is not a
and others (1985; Ainsworth, Blehar, Waters, & guarantee against pathology. A particularly import-
Wall, 1978), places the parent–child relationship ant finding from recent studies is that the form of
within an ethological, cognitive, control systems, and Insecure attachment in infants and young children
general developmental framework. It hypothesizes a termed ‘Insecure-Disorganized’ is strongly related to
biologically based need to form close affectional risk for psychopathology, compared with a Secure
bonds, and a specific developmental course. It fur- attachment and other more common forms of Inse-
ther proposes that the quality of the child’s experi- cure attachment, namely Avoidant and Ambivalent/
ences with his/her attachment figures (mother, Dependent (Green & Goldwyn, 2002; Greenberg,
father, or other caregiver) plays a central role in 1999; Lyons-Ruth, 1996). Following this observa-
personality and social development, and in the tion, interventions have targeted Disorganized
intergenerational transmission of caregiving quality. attachment as a focus for attention (e.g., Marvin,
In the context of a custody evaluation, several spe- Cooper, Hoffman, & Powell, 2002).
cific features of the theory are especially important. A fourth component of attachment theory with
First, attachment relationships provide the child particular relevance to the custody determination is
with protection against harm and with a sense of that early attachment relationships are internalized
emotional security, that is, a ‘secure base.’ A corol- and carried forward to influence expectations for
lary of this is that threats to or disruptions in the other important relationships. Bowlby referred to the
attachment relationships – such as that associated proposed mechanism mediating this process as an
with parental separation – would be expected to lead ‘internal working model,’ which can be described as
to fear/anxiety and accentuate attachment behavior an affective-cognitive schema or set. Through a his-
(e.g., need for proximity to attachment figures, tory of consistent and sensitive care with the parent,
especially among young children). Second, consist- the child develops a model of self and others as lov-
ent with the notion that attachment quality reflects able and loving/helpful that may make him/her
the history of care the child receives in the relation- comparatively more likely to cope with challenge and
ship, a child may have different forms of attachment stress (e.g., by relying on others for support or
relationships with each parent (Steele, Steele, & Fo- guidance). Internal working models are ‘working’
nagy, 1996). Therefore, in some situations, custody insofar as they are responsive to input from real-life
arrangements may involve the balancing of two se- experiences in the relationship; substantial changes
cure or two insecure relationships, and in others it in the quality of relationship would be expected to
may involve a balancing of a secure and an insecure lead to changes in the security of the attachment
attachment relationship. In any event, an assess- relationship. For example, given that parent–child
ment of the child’s attachment with both parents is relationships may suffer in the short term following a
needed. It is important to place this observation separation (see, Hetherington & Clingempeel, 1992),
alongside the proposal that children develop it might be expected that attachment quality might
attachment relationships with a limited number of also be temporarily altered.
caregivers and that there exists a hierarchy of Fifth, parental separation is a powerful challenge
attachment relationships. According to the theory, to the child’s attachment relationships with mother
the primary attachment relationship (which may or and father (Bowlby, 1973, 1980). An application to
not be the mother) is, in some biological and devel- the custody context is that joint physical custody
opmental sense, more ‘significant’ than other involving frequent moves between houses may itself
attachment relationships. be a major source of stress. Young children are not
Third, attachment theory proposes that the quality designed to cope with the separation from their
of care provided to the child, particularly sensitivity parents or with shared physical custody that in-
and responsiveness, leads to a Secure (optimal) or volves alternating between two separate homes.
Insecure (non-optimal) attachment relationship. What factors influence children’s ability to cope with
These experiences are associated with later devel- physical custody and visitation schedules are not
opment by setting in place adaptive or maladaptive well known, but the child’s age is one important
pathways. The term ‘pathways’ is used to make factor. Our clinical impression (empirical evidence is
explicit that early attachment experiences or, for that lacking) is that managing the stresses of frequent
matter, attachment experiences at any point in separations from parents that occur in joint physical
development, do not shape subsequent development custody are difficult for the infant and preschooler,
in a fixed, deterministic manner (Bowlby, 1988). A but may be more manageable by the school-aged
Practitioner Review: The contribution of attachment theory to child custody assessments 119

child. Whatever the effects on young children (a topic crawl to mother on her return to be picked up, the
addressed below), support for the possible positive secure preschool-aged child may make do with a
effects of living in two households for school-aged bright greeting at a distance and personal conver-
children was recently reported (Dunn, Davies, sation (Cassidy & Shaver, 1999; NICHD Early Child
O’Connor, & Sturgess, 2001). However, one paradox Care Research Network, 2001). An alternative to the
here is that many judges are loath to order move- lab-based assessment is the Attachment Q-sort, a
ment from one home to another after formal school- system that relies on observations of child and
ing has begun, but nevertheless order movement caregiver behavior in the home, typically over 2–3
from one home to another earlier in the child’s life, visits for a total of 2–6 hours’ observation (Pederson
when the child may be less able to cope with such a & Moran, 1996; Waters, Vaughan, Posada, & Kondo-
change. Ikemura, 1995).
Sixth, the potential contribution of attachment For children from approximately 3 years through
theory to the custody context is suggested by the age 8 years, measurement of the child’s representa-
many existing applications to clinical practice and tion or working model of attachment is increasingly
policy. For example, several groups have developed measurable through cognitive- and language-based
effective interventions specifically based on attach- assessments, such as story stem, doll play and
ment theory for a wide range of clinical problems related strategies. In these procedures, a child is
(Cicchetti, Rogosch, & Toth, 2000; Dozier, Stovall, shown toy figures and told the beginning of a story
Albus, & Bates, 2001; Steele, Hodges, Kaniuk, Hill- with an attachment theme (e.g., the child is hurt) and
man, & Henderson, 2003; for reviews, see Baker- is asked to complete the story; the child is prompted
mans-Kranenburg, van IJzendoorn, & Juffer, 2003; for information about how the child and parent fig-
Lieberman & Zeanah, 1999). ures would feel and what would happen next. The
Finally, it is important to note that attachment content of the child’s story is coded for instances of
does not encompass the whole of the parent–child secure, avoidant, dependent, aggressive, chaotic, or
relationship. There are many components of the other themes; in some cases the quality of his/her
parent–child relationship besides attachment, narrative is coded to reflect a coherent story with a
including teaching, disciplining, and caring for the clear beginning, middle, and end (resolution) versus
child’s physical needs, that also require careful a jumbled or incoherent style in which the ‘story line’
assessment. is hard to follow and make sense of (Bretherton,
Ridgeway, & Cassidy, 1990; Green, Goldwyn, &
Stanley, 2000; Oppenheim, Nir, Warren, & Emde,
Evidence-based measures of attachment
1997). The narrative procedures are relatively new
Research on attachment theory has led to the and their validation is modest, but they provide a
development of several standardized measures. useful window into the child’s cognitive set or
Some discussion of these measures is needed to expectations for the behavior of self and attachment
illustrate the concepts of attachment and to distin- figures. Compared with typical projective techniques
guish those measures of attachment that are sup- (e.g., drawings or apperception tests), these proced-
ported by scientific and clinical evidence. ures provide an empirically grounded and concep-
The most well-researched measure is the Strange tually sound method to conduct doll house/doll
Situation (Ainsworth et al., 1978), a structured story assessments for children of this age.
laboratory or clinic-based assessment in which the Until recently, late childhood and early adoles-
child is exposed to increasing stress, from play in an cence was an age period for which there was no
unfamiliar setting, to the introduction of a strange available measure of attachment. That is changing.
adult, to being left alone for a few minutes in the One recent example is the Child Attachment Inter-
strange setting. Although initially devised for 12–18- view (Humfress, O’Connor, Slaughter, Target, &
month-olds, the Strange Situation and separation– Fonagy, 2002; Target, Fonagy, & Shmueli-Goetz,
reunion procedures have been adapted for children 2003), an interview-based method in which the
into the preschool and early school age (Cassidy, child is asked to recall descriptors of relationship
Marvin, & the MacArthur Working Group on At- qualities and then asked to provide specific mem-
tachment, 1992; Crittenden, 1992; Main & Cassidy, ories that support those descriptors; questions
1988). A key feature of the coding is the child’s be- concerning instances of conflict and loss are also
havior upon the parent’s return following a separ- asked. Coding of the interview requires considera-
ation. Thus, having introduced a mild stress (i.e., tion of what the child says, but more importantly
separation), the assessment of attachment focuses how he/she describes the relationship episodes.
on how well the child uses the parent as a secure Narratives that are viewed as coherent and organ-
base. That is, how well is the child able to seek out ized are distinguished from those that are incoher-
the parent for comfort, emotional regulation, and ent, contradictory, minimalist, contain a poverty of
soothing, and then return to exploration? How the emotional language, and/or have themes of neglect,
child expresses these behaviors changes with strong rejection, violence or destruction without
development. Whereas a secure infant will likely successful resolution.
120 James G. Byrne et al.

A key feature of the attachment measures is that especially for assessing parent–child interactions,
they ‘activate’ the child’s attachment system (either have a weak empirical and conceptual base. We have
behaviorally or cognitively) by introducing mild proposed an attachment framework for child custody
stress in a standardized manner. The assessment evaluations based on its strong empirical foundation
strategy of inducing mild stress and examining how and conceptual relevance to the context of custody
the child uses or represents the parent as a secure evaluation. Further strengths could be added. For
base contrasts with the more typical assessment of example, reflecting the increased awareness of chil-
parent–child interactions in a low stress setting, dren’s rights, a number of international conventions
such as a clinic playroom or the home. Observing (UN Convention on the Rights of the Child; European
parent–child interactions only during a play or other Convention on the Exercise of Children’s Rights;
non/low-stress assessment may be irrelevant to Kovera & McAuliff 1999) require that the child’s
attachment (e.g., these observations are more relev- wishes are accounted for in matters affecting his/her
ant to parental support of the child’s play). It would welfare. This can present significant difficulties for
be inappropriate to make inferences about attach- custody evaluations. Direct examination of the
ment quality from such an assessment – although child’s wishes in a custody evaluation or, more
our impression is that this is not uncommon. broadly, the child’s-eye view of the relationships may
Measures of attachment quality in adults are also be exceptionally stressful, perhaps especially for
available and could be incorporated into a clinical younger (primary school-aged) children, for children
assessment as one predictor of the parent’s caregiv- with a very insecure attachment, or children who feel
ing strengths and risks, and as an aid to under- pressured to defend one or both parents. These
standing sources of a parent’s caregiving deficits. children have substantial difficulty knowing and
The most widely validated measure is the Adult reporting to outsiders what their wishes are.
Attachment Interview (AAI; Main & Goldwyn, 1998). Attachment assessments based for this age group
Adult attachment security as measured on the AAI is provide information about the child’s perspective
based on the adult’s ability to talk in a coherent, and experiences in a non-invasive and non-threat-
organized, and reflective manner when talking about ening manner. Furthermore, information gathered
past experiences with parents while growing up; as part of a standard attachment assessment of the
determining attachment security is less about what child–parent relationship may be less subject to the
the adult mentions experiencing in the course of the usual sort of bias that is common in custody evalu-
interview and more about how the adult talks about ations (Galatzer-Levy & Ostrov, 1999) and may even
these experiences, i.e., coherence of discourse. Of offer the opportunity to evaluate the likely validity of
particular interest to the clinical setting is the find- the parent’s perspective. Finally, it is worth consid-
ing that, for those adults who have experienced past ering how attachment theory may create and shape
trauma, the key consideration is not that a past psycho-educational opportunities for parents and
trauma was reported but rather how the adult talks the legal system. Specifically, an attachment evalu-
about these experiences and if s/he demonstrates ation can provide a context in which the parents and
some perspective on, or resolution of, these experi- court can be informed about how best to respond to
ences. One of the most important and robust find- the child(ren)’s emotional needs during and following
ings in the attachment literature is a high the separation proceedings (e.g., insuring availability
correspondence between the parent’s discussion of attachment figures) and how the separation pro-
during the AAI and the child’s attachment security cess may affect the child(ren) (e.g., anger and
assessed using the Strange Situation (van IJzendo- heightened separation worries).
orn, 1995). As a result of this evidence, use of the AAI Despite the ready acceptance of attachment theory
in clinic settings is gaining popularity. to the custody context (e.g., as illustrated by existing
The preceding review covers the most empirically guidelines), it is important to note several caveats
validated procedures for assessing child–parent and limitations. For instance, there are several
attachment. There is uncertainty about how well examples of how attachment theory has been seri-
these measures perform in the custody context, a ously misapplied to clinical problems. The most
topic to which we return later. The key take-home striking example is the use of attachment theory to
message is that if attachment is to be incorporated promote holding therapy for children with suspected
into the custody assessment, then there are clear ‘attachment disorders’ (Levy, 2000). An attachment
principles of measurement and that extant data theory perspective is, in fact, more antithetical than
support a restricted list of measures from which to congruent with the brand of ‘holding therapies’ pro-
choose. moted for these children. There is also a risk that
attachment concepts and measures would be mis-
applied and mis-communicated in the child custody
A profile of strengths and limitations of an
context. This is already happening. One recent
attachment theory model of custody evaluations
example is the analysis by Arredondo and Edwards
A central theme in this review is that measures most (2000), whose understanding of attachment theory
commonly used in child custody evaluations, was based on an amalgam of the research and
Practitioner Review: The contribution of attachment theory to child custody assessments 121

theoretical work of Bowlby and Ainsworth and the concern. Finally, whether or not these assessment
more recent interest in holding and related therapies approaches are more or less cost-effective than other
misleadingly referred to as attachment-based theor- approaches is not clear.
ies. As a result, their appraisal of attachment theory
as a tool for custody evaluations was predictably
negative. As we have noted, careful attention must Summary
be paid to how attachment is conceptualized and As a summary statement, we suggest two levels of
operationalized as part of an assessment. The aim integration of attachment theory in the custody
here is not to guard the use of the term ‘attachment’ evaluation context. The first is conceptual. The
for exclusive use by a certain minority, but instead to above discussion indicates that there is much to
point out that there are contradictions and incon- recommend about attachment theory as a concep-
sistencies in what is meant by ‘attachment,’ even tual framework for the practice of child custody
among health-care professionals. evaluations, and that this is already occurring.
A second set of limitations concerns the measure Attachment theory provides an interpretive frame-
of attachment. For example, existing measures of work for ‘making sense of’ the dynamics of children’s
attachment have not yet received scrutiny in the behavior that may be distressing and confusing to
context of divorce and custody litigation. We do not parents. For example, increased clinginess to par-
yet know how robust the information gathered from ents and anxiety about parents’ whereabouts and
the assessment is to the immediate stresses and safety make sense in the context of understanding
conflicts occurring in the family, however temporary the custody process as a threat to the child’s felt
they may be. Nor do we yet know how children’s re- sense of security. An additional illustration is that
sponses in the attachment-based assessments attachment theory suggests that the child’s anger
would be influenced by the context of custody toward the resident parent upon return from visiting
evaluation. Clinical experience shows that these the non-resident parent may reflect a natural re-
measures provide useful clinical information, but sponse to separation and loss (e.g., Bowlby, 1973) –
until systematic data are collated it is important to and not necessarily an undermining influence of the
re-state the potential concern of using measures for nonresident parent (although that does occur in
purposes other than those for which they were some cases).
developed. Indeed, there is likely to be disagreement A second level of integrating attachment in the
even among attachment researchers about the suit- child custody context is more practical and much
ability of these measures for clinical settings, such less easily answered: whether or not to use evidence-
as custody determination. In addition, questions based measures of attachment as part of the formal
have been raised about the stability of assessments, assessment. On the one hand, attachment research
even over comparatively short time periods. To some measures have been shown to be reliable and valid
extent, discontinuity may reflect real changes in the and among the best predictors of current and future
circumstances that impair or improve the parent’s social and behavioral development in the child.
sensitivity, but instability has been found even in the Thus, in any context, they would be preferable to the
absence of major disruptions to the parent–child largely unsupported and under-researched meas-
relationship (Belsky, Campbell, Cohn, & Moore, ures of parent–child relationship quality currently
1996). A related concern is that the prediction from used in custody evaluations. On the other hand,
one form of assessment to another (e.g., Strange there is not an evidence base to support their use
Situation to story stem method) is moderate rather specifically in the custody context. Our clinical
than strong. This may reflect ‘lawful’ discontinuity, experience indicates that these measures are useful
but may also mean that we have not yet identified the for characterizing the child–parent attachment rela-
age-based ‘equivalent’ manner of assessing child– tionship, identifying parenting strengths and weak-
parent attachment from infancy to adolescence. In nesses, and forecasting likely outcomes of
addition, there are practical limitations of applying alternative custody options. In any event, they war-
an attachment perspective. Chief among these is rant careful consideration. What is needed to in-
that some assessments require additional resources crease the evidence base of an attachment model of
(e.g., videotaping) and the coding requires extensive custody evaluation is a programmatic set of studies.
training that is costly and may be difficult to obtain. This is the topic of the final section.
Also, given that assessments of the child(ren)’s
attachment with both parents is needed, the
procedures may become cumbersome, especially if
4. Practical application: translating theory into
there are several children involved. Concerns about
research to inform practice
repeating the attachment assessment with the child
for each parent (e.g., is the situation as ‘strange’ on In this final section we outline some of the kinds of
the repeat assessment?) might also be raised, al- research that could move the inclusion of attach-
though if the visits are conducted on different days ment theory in child custody determination from
(as is usual practice), this is not likely to be a major theoretical appeal to an evidence-based decision.
122 James G. Byrne et al.

Before sketching out a research agenda, some efforts be focused on a) providing independent sup-
general points about this line of research deserve port for the measures when used in custody evalu-
mention. ation, b) evaluating the extent to which attachment
A positive feature of this research agenda is that it theory and measures enhance our understanding of
breaks down the (largely artificial) distinction be- what predicts optimal outcomes, and c) broadening
tween ‘basic’ and ‘applied’ research, and toward a the focus of attachment to address un-addressed
situation in which there is a genuine dialectic be- questions and links with existing research on the
tween theory and clinical practice (e.g., Toth, Mau- post-divorce family.
ghan, Manly, Spagnola, & Cicchetti, 2002). On the We expand on these three tasks for research
other hand, research into custody evaluation is below.
complicated. One major challenge is that the ‘real The first task for research concerns the use of
world’ context in which research on custody would attachment measures in the custody context: is
take place means that we need to balance the de- information gathered from attachment measures as
mand of scientific rigor with the constraints of what valid and reliable when collected as part of a custody
is ethically and practically possible. Ordinary re- evaluation as during other clinical contexts? One
search concerns about, for example, experimental approach to this task is to determine if the construct
design and random assignment cannot easily be validity of the measures is challenged. This could be
met; instead, this area is better approached as a assessed in several ways. For example, is there the
‘natural experiment.’ ‘Real world’ challenges come in ordinarily high concordance between results from
several forms. For instance, at the time the family the child’s Strange Situation and results from the
presents for assessment and likely for research, they parent’s Adult Attachment Interview when this
are in an intense period of distress (e.g., initial non- information is collected during the custody evalua-
adversarial strategies have failed) and there is likely tion? This connection has been proposed by some to
to be a de facto custody arrangement in place (e.g., be among the strongest demonstrations of construct
the child’s contact with the father may already be validity of the measures; if this condition is not met
inconsistent by the time the custody determination when data are collected in the course of a custody
begins). Clearly, it would be better to start following assessment, then that would be a major source of
families prior to this crisis period for both scientific concern. In addition, do the patterns of attachment
reasons and ethical and practical concerns about behavior and classifications observed in custody
getting consent. That may be possible where close assessments conform to those found in a range of
links exist with family law courts and therapists in normal and high-risk samples studied to date; or, is
the community. Another consideration is that cus- there an increased rate of ‘unclassifiable’ and atyp-
tody placement at the point of separation is better ical behaviors as judged by an independent coder? If
viewed as a fluid than final decision; children regu- the latter is the case, then that would also be a
larly change households, and anecdotal evidence source of concern about the measures. Moreover, if,
suggests that few of these changes are prompted by as some have hypothesized, the custody context may
or even known to the courts. be so threatening to the child as to increase rates of
A further difficulty is determining what good out- Disorganized attachment, then we would expect
comes measures are for research. We suggest that elevated rates of Disorganization in those undergo-
custody determination is not a good outcome for re- ing evaluation (accounting for the likelihood that
search purposes and that judicial decision is not a Disorganized attachment is associated with co-
useful criterion against which to evaluate the use- occurring risks that may also be evident in families
fulness or validity of measures used in a custody requiring psychological consultation, such as mal-
evaluation. As noted earlier, many factors determine treatment or serious parental mental illness). If there
custody decisions; different judges – even in the is an elevated rate of Disorganization in children
same jurisdiction – may weight other information as when assessed in a custody evaluation, then the
more relevant, or accepted practice may bias against clinician would need to be more wary about inter-
an open consideration of the evidence presented preting the Disorganization as an index of
(e.g., a value of joint custody may be viewed as more heightened risk for maladjustment. That is, if Dis-
compelling than results from psychological evalua- organization is ‘exaggerated’ by the immediate
tions). Other variables might be substituted as out- stresses of the peri-separation and custody context,
comes, such as judges’ reports of how helpful or it may not be as predictive of child adjustment as it is
influential the psychological assessment was to the for children who are classified as Disorganized in
custody evaluation. Information of that kind may be more routine community and clinical research set-
important, but might well be regarded as the wrong tings. A further test of the integrity of attachment
focus of attention – i.e., it is the well-being of child measures is whether the information obtained in the
and parents that is of central importance. Moreover, course of a custody evaluation is reliable, using
experience tells us that there is wide variation among conventional criteria such as test–retest reliability.
judges in the value placed on psychological assess- Findings from the sorts of studies outlined above
ment material. Instead, we suggest that research cannot ‘rule in’ the appropriateness of attachment
Practitioner Review: The contribution of attachment theory to child custody assessments 123

measures, but would represent a first test of whether In fact, there is currently a great deal of social and
attachment data have comparable meaning when scientific controversy concerning joint physical cus-
gathered in a custody evaluation as during other tody, the amount of visitation that is optimal or
periods. minimal to maintain attachment relationships with
A second task which bears directly on the ‘best the non-resident parent, the conditions under which
interests’ standard is to determine how predictive an attachment to non-resident parent should be fos-
attachment model is of post-separation/decision tered – even in the face of ongoing and severe par-
adjustment. This kind of research capitalizes on the ental conflict, the effects on the child of stopping
previously mentioned observation that judicial contact between him/her and a parent, and whether
decision will, with some regularity, contradict rec- relocation of one parent should be granted if it
ommendations made on the basis of attachment means that contact with the other parent would be
theory and data. The point here is not to test the severely affected (see, Gould & Stahl, 2001; Kelly &
hypothesis that attachment data should be the sole Lamb, 2000, 2003; Solomon & Beringen, 2001).
deciding factor in custody determination (it should Decisions on these matters are part of most custody
not be, as we have stated). Rather, the aim of this cases involving psychological input; unfortunately,
sort of research is to get an empirical hold on the for none of the above topics is there adequate evid-
significance of other competing or complementary ence guiding decisions one way or another. Whether
factors – from parental work situations to biases joint physical custody should be granted for parents
favoring maternal or joint custody – for optimal of infants and young children is a case in point. On
custody outcome, and what sort of balance among the basis of research findings pointing to the fact
competing factors may be most appropriate. This that infants form attachment with both parents and
type of research is more intensive insofar as it re- that young children learn to manage transitioning
quires longitudinal follow-up into the post-divorce between attachment figures (e.g., in the case of
period, with detailed assessments that would not be childcare providers), Kelly and Lamb (2000) rea-
part of the custody evaluation. Studies that did col- soned that joint physical custody for infants should
lect that kind of data would be in a position to be encouraged, alongside complementary efforts to
determine if children whose custodial placement facilitate co-parenting between separated parents;
contrasts with the attachment data showed poorer they also noted that failing to support the infant’s
outcomes compared with children whose custody attachment to one parent would be harmful. In
placement was congruent with the attachment contrast, Solomon and Beringen (2001) espoused the
assessment information. Gathering attachment data opposite view, based on the findings of Solomon and
in the course of the custody evaluation would also George (2001) and the observation that maintaining
allow clinical investigators to examine if children a stable setting should be prioritized for young chil-
with Insecure attachment patterns (perhaps especi- dren. The fact that opposite conclusions can be
ally Disorganized) are more likely to show a greater reached based on attachment theory and available
relative increase in behavioral/emotional problems research illustrates how critical the need is for di-
in the course of the custody process and into the rectly relevant, rigorous research. Whether opposing
post-separation period, as compared with children views would be proposed for older children is not
with attachment relationship(s) judged Secure. clear, but research on older children is just as
Whether or not attachment data are included in important given the regularity with which school-age
the custody evaluation, attachment data may pro- and adolescent children are forced to cope with
vide valuable insights into the effects of custody separating parents.
determination. One of the rare studies of this type A third task for research is to expand the bound-
was reported by Solomon and George (2001). They aries of attachment theory by exploring questions
found that infants who had overnight visits with both that would be otherwise difficult to address and by
divorced parents (i.e., some degree of joint physical linking attachment concepts in to the substantial set
custody) were more likely to show Disorganized of studies on children’s adjustment in post-divorce
attachments, compared with infants who did not families (Amato & Gilbreth, 1999; Bauserman, 2002;
have overnight visits with the non-resident father Buchanan, Maccoby, & Dornbusch, 1996; Dunn,
and infants in non-divorced families. This was Davies, O’Connor, & Sturgess, in press; Gunnoe &
especially the case when there was parental conflict Braver, 2001). The first kind of expansion capitalizes
over custody and/or visitation. Of course, the higher on divorce and the custody context as a natural ex-
rate of Disorganized attachment in the overnight periment to address novel questions. One such op-
visitation group could be attributed to the increased portunity would be to examine the concept of
stress of repeated separations, conflict, and com- attachment to the broader family system. Several
promised parenting that preceded the divorce. theorists have suggested that attachment is not only
Therefore, it cannot be concluded if the visitation a dyadic concept, but also illuminates family sys-
arrangement had any causal role in the child’s Dis- tems processes (Byng-Hall, 1999; Hill, Fonagy,
organized attachment. The findings require replica- Safier, & Sargent, 2003; Marvin & Stewart, 1990;
tion and further consideration. Wynne, 1984). In the context of custody and
124 James G. Byrne et al.

separation, the clinical and research question is, communicating findings and recommendations to
what is the effect on the child of a disrupted parents and the legal system.
attachment to the family? Thus, even a child who
maintained secure attachment relationships with
both parents during the separation process would Conclusion
undergo ‘attachment stress’; that is, the child’s sense
Practice guidelines from professional bodies and
of safety and security might well be threatened de-
numerous books and articles do not reflect, and have
spite having secure attachment with both parents.
not led to, evidence-based practice for child custody
Davies and Cummings’ (1984) model for researching
determination. In this review we tackled one way of
the child’s emotional security in the context of mar-
improving the fit between science and practice in this
ital conflict may provide one empirical strategy for
area, namely, by describing how attachment theory
consolidating dyadic attachment and family systems
might be applied in custody evaluations. Specific-
approaches to understanding the psychological
ally, we articulated why attachment theory may be of
dynamics and effects of parental separation.
particular value, described the reliable and valid
Similarly, there are opportunities to broaden
measures that have been used to index attachment
attachment research by linking it with other lines of
processes, and outlined the kinds of evidence that
research on children in post-divorce families. For
are (still) needed to develop an attachment-oriented
example, data from several countries show that
evidence-based model for practice.
children’s contact with the non-resident father drops
substantially during the period following the
separation (Maclean & Eekelaar, 1997; Pryor &
Correspondence to
Rodgers, 2001; Seltzer, 1991). Research has investi-
gated the non-resident parent–child relationship in Gerard Byrne, Lucena Clinic, Blessington Road,
terms of the predictors of contact, the role of the Tallaght, Dublin 24, Ireland, or Tom O’Connor,
resident parent, and the impact of contact as distinct Department of Psychiatry, University of Rochester
from quality of relationship (Amato & Gilbreth, 1999; Medical Center, 300 Crittenden Blvd, Rochester, NY,
Dunn et al., in press). However, the child’s attach- 14642, USA; Email tom_oconnor@urmc.rochester.
ment relationship with the non-resident parent has edu
been neglected. An attachment perspective would
pose novel and important questions for this line of
research, such as what is the effect of the reduction in References
contact/loss of relationship on the child? Does the
Ackerman, M.J., & Ackerman, M.C. (1997). Custody
non-resident parent cease to serve as an attachment evaluation practices: A survey of experienced profes-
figure even if there is regular (albeit) infrequent con- sionals (revisited). Professional Psychology: Research
tact, such as one weekend a month? How does the and Practice, 28, 137–145.
child’s developmental level modify the amount of Ackerman, M.J., & Schoendorf, K. (1992). ASPECT:
contact that is optimal for maintaining the attach- Ackerman–Schoendorf Scales for Parent Evaluation of
ment relationship and the desirability of infrequent Custody. Los Angeles, CA: Western Psychological
visitation? Numerous questions that regularly arise Services.
in clinical practice now need to be subjected to Ainsworth, M.D.S. (1985). Attachments across the life
clinical research. span. Bulletin of the New York Academy of Medicine,
61, 792–812.
Many pressing and perplexing questions confront
Ainsworth, M.D.S., Blehar, M.C., Waters, E., & Wall, S.
the clinician involved in custody assessments:
(1978). Patterns of attachment: A psychological study
should a child be placed with the primary caregiver of the Strange Situation. Hillsdale, NJ: Erlbaum.
even if the primary caregiver is judged to be a less Amato, P.R., & Gilbreth, J.G. (1999). Nonresident
competent parent than the non-primary caregiver? Is fathers and children’s well-being: A meta-analysis.
joint physical custody to be recommended if the in- Journal of Marriage and the Family, 61, 557–573.
fant child shows Secure attachment to both parents American Academy of Child and Adolescent Psychiatry.
and both parents share the primary caregiving role? (1997). Practice parameters for child custody evalu-
Not all of these questions – as they pertain to the ation. Journal of the American Academy of Child and
individual case – can be firmly answered by clinical Adolescent Psychiatry, 36, 57s–68s.
experience or research findings, and there is certain American Psychological Association. (1994). Guidelines
for child custody evaluations in divorce proceedings.
to be dissent among clinicians, even those know-
American Psychologist, 49, 677–680.
ledgeable about attachment. Ultimately, therefore,
Arredondo, D.E., & Edwards, L.P. (2000). Attachment,
the incorporation of attachment principles and bonding, and reciprocal connectedness: Limitations
measures into practice will also depend not on its of attachment theory in the Juvenile and Family
ability to provide fool-proof answers, but on whether Courts. Journal of the Center for Families, Children,
the clinician perceives insights into the case that and the Courts, 2, 109–127.
might otherwise be difficult to obtain and whether an Bakermans-Kranenburg, M.J., van IJzendoorn, M.H., &
attachment perspective offers a ‘user-friendly’ way of Juffer, F. (2003). Less is more: Meta-analysis of
Practitioner Review: The contribution of attachment theory to child custody assessments 125

sensitivity and attachment interventions in early fathers: Influences, outcomes, and implications. Jour-
childhood. Psychological Bulletin, 129, 195–215. nal of Child Psychology and Psychiatry, 45, 553–566.
Bartlett, K.T. (1999). Improving the law relating to Dunn, J., Davies, L., O’Connor, T., & Sturgess, W.
postdivorce arrangements for children. In R.A. (2001). Family lives and friendships: The perspectives
Thompson & P.R. Amato (Eds.), The postdivorce of children in step-, single-parent and nonstep
family (pp. 71–102). Thousand Oaks, CA: Sage. families. Journal of Family Psychology, 15, 272–287.
Bauserman, R. (2002). Child adjustment in joint-cus- Galatzer-Levy, R.M., & Ostrov, E. (1999). From empiri-
tody versus sole-custody arrangements: A meta-ana- cal findings to custody decisions. In R.M. Galatzer-
lytic review. Journal of Family Psychology, 16, Levy & L. Kraus (Eds.), The scientific basis of child
91–102. custody decisions, (pp. 32–57). New York: Wiley.
Belsky, J., Campbell, S., Cohn, J., & Moore, G. (1996). Gerard, A.B. (1994). Parent–Child Relationship Inven-
Instability of attachment security. Developmental tory (PCRI): Manual. Los Angeles, CA: Western Psy-
Psychology, 32, 921–924. chological Services.
Bernet, W. (2002). Child custody evaluations. Child and Goldstein, J., Freud, A., & Solnit, A.J. (1980). Beyond
Adolescent Psychiatry Clinics of North America, 11, the best interests of the child. London: Burnett Books.
781–804. Gould, J.W., & Stahl, P.M. (2001). Never paint by
Bowlby, J. (1973). Attachment and loss: Vol. II Separa- numbers: A response to Kelly & Lamb (2000),
tion: Anxiety and anger. New York: Basic Books. Solomon (2001), and Lamb & Kelly (2001). Family
Bowlby, J. (1980). Attachment and loss: Vol. III. Loss: Court Review, 39, 372–376.
Sadness and depression. New York: Basic Books. Green, J., & Goldwyn, R. (2002). Annotation: Attach-
Bowlby, J. (1982). Attachment and loss: Vol. I. Attach- ment disorganization and psychopathology: New
ment (2nd edn). New York: Basic Books. findings in attachment research and their practical
Bowlby, J. (1988). Developmental psychiatry comes of implications for developmental psychopathology in
age. American Journal of Psychiatry, 145, 1–10. childhood. Journal of Child Psychology and Psychi-
Bretherton, I., Ridgeway, D., & Cassidy, J. (1990). atry, 43, 835–846.
Assessing internal working models of the attachment Green, J., Goldwyn, R., & Stanley, C. (2000). A new
relationship: An attachment story completion task for method of evaluating attachment representations in
3-year-olds. In M.T. Greenbergh, D. Cicchetti, & E.M. young school-age children: The Manchester Child
Cummings (Eds.), Attachment in the preschool years Attachment Story Task. Attachment and Human
(pp. 273–308). Chicago: University of Chicago Press. Development, 2, 48–70.
Bricklin, B. (1990). PORT handbook: Perception-of-rela- Greenberg, M.T. (1999). Attachment and psychopathol-
tionships test. Furlong, PA: Village. ogy in childhood. In J. Cassidy & P. Shaver (Eds.),
Buchanan, C.M., Maccoby, E.E., & Dornbusch, S.M. Handbook of attachment (pp. 469–496). New York:
(1996). Adolescents after divorce. Cambridge, MA: Guilford Press.
Harvard University Press. Gunnoe, M.L., & Braver, S.L. (2001). The effects of joint
Byng-Hall, J. (1999). Family and couple therapy: legal custody on mothers, fathers, and children
Towards greater security. In J. Cassidy & P.R. Shaver controlling for factors that predispose a sole maternal
(Eds.), Handbook of attachment (pp. 625–645). New joint versus legal award. Law and Human Behavior,
York: Guilford Press. 25, 25–43.
Cassidy, J., Marvin, R.S., & The MacArthur Working Halikias, W. (1994). Forensic family evaluations: A
Group on Attachment (1992). Attachment organiza- comprehensive model for professional practice. Jour-
tion in preschool children: Coding guidelines (4th edn). nal of Clinical Psychology, 50, 951–964.
Unpublished manuscript, University of Virginia. Heinze, M., & Grisso, T. (1996). Review of instruments
Cassidy, J., & Shaver, P.R. (Eds.) (1999). Handbook of assessing parenting competencies used in child
attachment. New York: Guilford. custody evaluations. Behavioral Sciences and the
Cicchetti, D., Rogosch, F.A., & Toth, S.L. (2000). The Law, 14, 293–313.
efficacy of toddler–parent psychotherapy for fostering Hetherington, E.M., & Clingempeel, W.G. (1992). Cop-
cognitive development in offspring of depressed ing with marital transitions: A family systems per-
mothers. Journal of Abnormal Child Psychology, 28, spective. Monographs of the Society for Research in
135–148. Child Development, 57 (2–3, Serial no. 227).
Crittenden, P.M. (1992). Quality of attachment in the Hill, J., Fonagy, P., Safier, E., & Sargent, J. (2003). The
preschool years. Development and Psychopathology, ecology of attachment in the family. Family Process,
4, 209–241. 42, 205–221.
Davies, P.T., & Cummings, E.M. (1984). Marital conflict Humfress, H., O’Connor, T.G., Slaughter, J., Target, M.,
and child adjustment: An emotional security hypo- & Fonagy, P. (2002). Generalised and relationship-
thesis. Psychological Bulletin, 116, 387–411. specific models of social cognition: Explaining the
Derdeyn, A.P., & Scott, E. (1984). Joint custody: A overlap and discrepancies. Journal of Child Psycho-
critical analysis and appraisal. American Journal of logy and Psychiatry, 43, 873–883.
Orthopsychiatry, 54, 199–209. Hysjulien, C., Wood, B., & Benjamin, G.A.H. (1994).
Dozier, M., Stovall, K.C., Albus, K.E., & Bates, B. Child custody evaluations. A review of methods used
(2001). Attachment for infants in foster care: The role in litigation and alternative dispute resolution.
of caregiver state of mind. Child Development, 72, Family and Conciliation Courts Review, 32, 466–489.
1467–1477. Jameson, B.J., Ehrenberg, M.F., & Hunter, M.A. (1997).
Dunn, J., Cheng, H., O’Connor, T.G., & Bridges, L. (2004). Psychologists’ ratings of the best interests of the child
Children’s relationships with their non-resident custody and access criterion: A family systems
126 James G. Byrne et al.

assessment model. Professional Psychology: Re- Marvin, R.S., & Stewart, R.B. (1990). A family systems
search and Practice, 28, 253–262. framework for the study of attachment. In M.T.
Johnston, J.R., & Campbell, L.E. (1993). A clinical Greenberg, D. Cicchetti, & E.M. Cummings (Eds.).
typology of interparental violence in disputed-custody Attachment in the preschool years (pp. 51–86). Chi-
divorces. American Journal of Orthopsychiatry, 63, cago: The University of Chicago Press.
190–199. McDermott, J.F., Tseng, W.S., Char, W.S., & Fukunaga,
Keilin, W.G., & Bloom, L.J. (1986). Child custody C.S. (1978). Child custody decision making: The
evaluation practices: A survey of experienced profes- search for improvement. Journal of the American
sionals. Professional Psychology: Research and Prac- Academy of Child Psychiatry, 17, 104–116.
tice, 17, 338–346. Melton, G.B., Petrika, J., Poythress, N.G., & Slobogin,
Kelly, J.B. (1994). The determination of child custody. C. (1987). Psychological evaluations for the courts.
The Future of Children, 4, 121–142. New York: Guilford Press.
Kelly, J.B., & Lamb, M.E. (2000). Using child develop- NICHD Early Child Care Research Network (2001).
ment research to make appropriate custody recom- Child-care and family predictors of preschool attach-
mendations and access decisions for young children. ment and stability from infancy. Developmental
Family and Conciliation Courts Review, 38, 297–311. Psychology, 37, 847–862.
Kelly, J.B., & Lamb, M.E. (2003). Developmental issues O’Connor, T.G. (2002). Annotation: The ‘effects’ of
in relocation cases involving young children: When, parenting reconsidered: Findings, challenges, and
whether, and how? Journal of Family Psychology, 17, applications. Journal of Child Psychology and Psychi-
193–205. atry, 43, 555–572.
Klein, R.G., (1986). Questioning the clinical usefulness O’Donohue, W., & Bradley, A.R. (1999). Conceptual and
of projective psychological tests for children. Devel- empirical issues in child custody evaluations. Clinical
opmental and Behavioral Pediatrics, 7, 378–382. Psychology: Science and Practice, 6, 310–322.
Kovera, M. & McAuliff, B.D. (1999). Child witnesses in Oppenheim, D., Nir, A., Warren, S., & Emde, R.N.
custody cases: The effects of system and estimator (1997). Emotion regulation in mother–child narrative
variables on the accuracy of their reports. In R.M. co-construction: Associations with children’s narrat-
Galatzer-Levy & L. Kraus (Eds.), The scientific basis of ives and adaptation. Developmental Psychology, 33,
child custody decisions (pp. 157–187). New York: 284–294.
Wiley. Pederson, D.R., & Moran, G. (1996). Expressions of the
Kunin, C.C., Ebbesen, E.B., & Konecni, V.J. (1992). An attachment relationship outside of the strange situa-
archival study of decision-making in child custody tion. Child Development, 67, 915–927.
disputes. Journal of Clinical Psychology, 48, 564–573. Pryor, J., & Rodgers, B. (2001). Children in changing
Levy, T.M. (2000). (ed.) Handbook of attachment inter- families: Life after parental separation. Oxford, UK:
ventions. New York: Academic Press. Blackwell Publishers.
Lieberman, A.F., & Zeanah, C.H. (1999). Contributions Quinnell, F.A., & Bow, J.N. (2001). Psychological tests
of attachment theory to infant–parent psychotherapy used in child custody evaluations. Behavioral
and other interventions with infants and young Sciences and the Law, 19, 491–501.
children. In J. Cassidy & P. Shaver (Eds.), Handbook Ramchandani, P., Joughlin, C., & Zwi, M. (2001).
of attachment (pp. 555–574). New York: Guilford Evidence-based child and adolescent mental health
Press. services: Oxymoron or brave new dawn? Child
Lyons-Ruth, K. (1996). Attachment relationships Psychology and Psychiatry Review, 6, 59–64.
among children with aggressive behavior problems: Seltzer, J.A. (1991). Relationships between fathers and
The role of disorganized early attachment patterns. children who live apart: The father’s role after separa-
Journal of Clinical and Consulting Psychology, 64, 64– tion. Journal of Marriage and the Family, 53, 79–
73. 101.
Maccoby, E.E., & Mnookin, R.H. (1992). Dividing the Solomon, J. & Beringen, Z. (2001). Another look at the
child: Social and legal dilemmas of custody. Cam- developmental research: Commentary on Kelly and
bridge, MA: Harvard University Press. Lamb’s ‘using child developmental research to make
Maclean, M., & Eekelaar, J. (1997). The parental appropriate custody and access decisions’. Family
obligation: A study of parenthood across households. Court Review, 39, 355–364.
Oxford: Hart Publishing. Solomon, J., & George, C. (2001). The development of
Main, M., & Cassidy, J. (1988). Categories of response attachment in separated and divorced families:
to reunion with the parent at age six: Predictable from Effects of overnight visitation, parent and couple
infant attachment classifications and stale over a variables. Attachment and Human Development, 1,
one-month period. Developmental Psychology, 24, 2–33.
415–426. Steele, M., Hodges, J., Kaniuk, J., Hillman, S., &
Main, M., & Goldwyn, R. (1998). Adult attachment Henderson, K. (2003). Attachment representations
rating and classification systems. Unpublished and adoption: Associations between maternal states
manuscript, University of California, Berkeley, of mind and emotion narratives in previously mal-
Department of Psychology. treated children. Journal of Child Psychotherapy, 29,
Marvin, R.S., Cooper, G., Hoffman, K., & Powell, B. 187–205.
(2002). The circle of security project: Attachment- Steele, H., Steele, M., & Fonagy, P. (1996). Associations
based interventions with caregiver–pre-school child among attachment classifications of mothers,
dyads. Attachment and Human Development, 4, 107– fathers, and their infants. Child Development, 67,
124. 541–555.
Practitioner Review: The contribution of attachment theory to child custody assessments 127

Target, M., Fonagy, P., & Shmueli-Goetz, Y. (2003). ity of the Adult Attachment Interview. Psychological
Attachment representations in school-age children: Bulletin, 117, 387–403.
The development of the child attachment interview Waters, E., Vaughan, B., Posada, G., & Kondo-Ikemura,
(CAI). Journal of Child Psychotherapy, 29, 171–186. K. (1995). Caregiving, cultural, and cognitive per-
Toth, S.L., Maughan, A., Manly, J.T., Spagnola, M., & spectives on secure-base behavior and working
Cicchetti, D. (2002). The relative efficacy of two models: New growing points in attachment theory
interventions in altering maltreated preschool chil- and research. Monographs for the Society for Re-
dren’s representational models: Implications for search in Child Development, 60 (2–3, serial No. 244).
attachment theory. Development and Psychopathol- Wynne, L.C. (1984). The epigenesis of relational sys-
ogy, 14, 877–908. tems: A model for understanding family development.
van IJzendoorn, M.H. (1995). Adult attachment repre- Family Process, 23, 297–318.
sentations, parental responsiveness, and infant
attachment: A meta-analysis on the predictive valid- Manuscript accepted 1 July 2004

Das könnte Ihnen auch gefallen