Author's Accepted Manuscript

Emotion regulation as a mediator in the

relationship between attachment and depres-
sive symptomatology: A systematic review
Sonia Malik, Adrian Wells, Anja Wittkowski

www.elsevier.com/locate/jad

PII: S0165-0327(14)00629-6
DOI: http://dx.doi.org/10.1016/j.jad.2014.10.007
Reference: JAD7028

To appear in: Journal of Affective Disorders

Received date: 22 May 2014

Revised date: 1 October 2014
Accepted date: 3 October 2014

Cite this article as: Sonia Malik, Adrian Wells, Anja Wittkowski, Emotion
regulation as a mediator in the relationship between attachment and
depressive symptomatology: A systematic review, Journal of Affective Disorders,
http://dx.doi.org/10.1016/j.jad.2014.10.007

This is a PDF file of an unedited manuscript that has been accepted for
publication. As a service to our customers we are providing this early version of
the manuscript. The manuscript will undergo copyediting, typesetting, and
review of the resulting galley proof before it is published in its final citable form.
Please note that during the production process errors may be discovered which
could affect the content, and all legal disclaimers that apply to the journal
pertain.
 Emotion Regulation as a Mediator in the Relationship
between Attachment and Depressive Symptomatology:
A Systematic Review

Sonia Malik1, Adrian Wells1,2 & Anja Wittkowski1,2*

1
University of Manchester, School of Psychological Sciences, Manchester, United Kingdom
Manchester Mental Health and Social Care Trust, Manchester, United Kingdom

Word count (main text): 9,615

*Corresponding author: Dr Anja Wittkowski, School of Psychological Sciences, Zochonis

Building, Brunswick Street, The University of Manchester, Manchester, United Kingdom, M13
9PL. Phone: +44(0)161 306 0400; Fax: +44(0)161 306 0406; Email:
anja.wittkowski@manchester.ac.uk
Abstract

Background: Attachment theory has been conceptualised as an emotion regulation theory.

Research attributes the occurrence of depressive symptoms to a dysfunction of emotion regulation.
Anxious attachment and avoidant attachment, which are two dimensions of insecure attachment,
are hypothesised to lead to the development of hyperactivating and deactivating emotion
regulation strategies.
Methods: This systematic review examines the literature on the role of emotion regulation and its
relationship with attachment and depressive symptomatology. Furthermore, we examined evidence
for hyperactivating and deactivating strategies.
Results: Nineteen papers were identified. Adolescent studies demonstrated associations of varying
strength and found unreliable and contradictory results for emotion regulation as a mediator.
Conversely, adult studies provided strong evidence for emotion regulation as a mediator. The
hypothesis that hyperactivating strategies mediate anxious attachment and depressive symptoms
was consistently supported. Mixed evidence was provided for deactivating strategies as mediators
to avoidant attachment and depressive symptomatology.
Limitations: Limitations of methodology and quality of studies are identified with particular
attention drawn to problems with conceptual singularity and multicollinearity.
Conclusions: Despite mixed variable findings, this review indicates that emotion regulation is a
mediator between attachment and depression. Hyperactivating strategies, in particular, have been
consistently noted as mediators for anxious attachment and depressive symptomatology, whereas
evidence for deactivating strategies as mediators between avoidant attachment and depressive
symptoms has been mixed. Future research should test the mediators of attachment and symptoms
and examine theoretically grounded models of psychopathology, such as metacognitive and
cognitive models using clinical samples.

Keywords: Attachment, emotion regulation, mediator, depression

1
Introduction
Attachment theory (Bowlby, 1973, 1980, 1982) stipulates that an innate range of behaviours aim at
establishing proximity to caregivers, which are considered attachment figures. These attachment
behaviours are hypothesised to regulate affect, alleviate distress and protect from physical and
emotional threats (Bowlby, 1988; Mikulincer et al., 2003). Attachment behaviours are part of an
attachment behavioural system, which extends from initial primary caregivers to partners in adult
relationships (Mikulincer et al., 2003).
Studies have shown that attachment is associated with depressive disorders (Besser and
Priel, 2003; Carnelley et al., 1994; Roberts et al., 1996; Wei et al., 2003; Wei et al., 2004; Wei et
al., 2006). As attachment theory has been frequently conceptualised as an emotion regulation
theory (Feeney and Noller, 1996; Mikulincer et al., 2003), it is surprising that there are limited
studies to test emotion regulation as mediator in the relationship between attachment and
depression. Such research would have implications for the development of clinical interventions
targeting mediators of psychological distress, as opposed to more resource intensive interventions
aimed at altering attachment styles (Wei et al., 2003). Thus, a systematic review of the literature
on the relationship between attachment, emotion regulation and depressive symptomatology is
warranted. Furthermore, findings in this area may provide clinicians with further insight into the
mediators to insecure attachment and depressive symptomatology, whilst also highlighting areas of
future research in this area.
Whilst there is a review on emotion regulation as a mediator for attachment and anxiety
(Esbjrn et al., 2012), the authors are not aware of any reviews on the associations between
attachment, emotion regulation and depressive symptomatology. Thus, the aim of the current
systematic review was to synthesise literature on the relationships between attachment, emotion
regulation and depressive symptomatology and to explore emotion regulation as mediator in the
relationship between insecure attachment and depressive symptoms. Furthermore, the relationships
between specific dimensions of insecure attachment and types of emotion regulation on depressive
symptomatology were explored (see Figure A.1 for our hypothesised model of the theoretical links
between attachment, emotion regulation and depressive symptoms). This review was aimed at
aiding understanding of how attachment, emotion regulation and depressive symptomatology may
be related and the reliability of such associations in order to provide a basis for linking attachment
theory to psychopathology and to suggest directions for future research.
2
 [Insert Figure A.1 about here]

Attachment Orientations
Insecure attachment is described using two dimensions, anxious attachment and avoidant
attachment (Bartholomew and Horowitz, 1991; Brennan et al., 1998). A securely attached
individual will present as low on anxiety and low on avoidance.
An individual with anxious attachment is hypothesised to use hyperactivation strategies
(Main, 1990; Collins et al., 2006) in response to an inconsistent caregiver (Ainsworth et al. 1978).
Hyperactivation strategies involve demanding proximity and care from attachment figures
(Mikulincer et al., 2003) including hyper-vigilance towards attachment cues, the use of worry, and
rumination (Cassidy, 1994). These strategies often fail to regulate emotions and can amplify
distress (Mikulincer and Shaver, 2007).
Avoidant attachment is characterised by compulsive self-reliance and the use of
deactivating strategies (Cassidy and Kobak, 1988), following an appraisal that proximity seeking
behaviours are unfruitful. Such strategies involve creating an emotional distance from others
(Mikulincer et al., 2003), which is hypothesised to extend to an inattention to personal
vulnerabilities and threatening incidences. Other deactivating strategies include the suppression of
distressing cognitions and memories (Shaver and Mikulincer, 2002). Hyperactivating and
deactivating strategies are secondary attachment strategies following the failure of distress
relieving responses from attachment figures.
Classifications of insecure attachment styles have been used in adult attachment literature
(Bartholomew and Horowitz, 1991), such as preoccupied attachment which is characterised by
high anxiety and low avoidance, dismissing-avoidant attachment, characterised by high avoidance
and low anxiety, and fearful-avoidant attachment style which presents as high anxiety and high
avoidance. For the purposes of clarity within this review, we chose to use the terms secure,
anxious (to represent anxious and preoccupied attachment) and avoidant (to represent avoidant and
dismissing attachment).
Fraley and Waller (1998) recommended the use of continuous scoring for attachment
measures rather than categorical assessments due to limitations arising from the use of attachment
categories, such as the description of individuals using mixed attachment types and difficulty in
3
assessing the stability of attachment orientations. Furthermore, confirmatory factor analyses
(Kurdek, 2002; Roisman et al., 2007) indicate the use of the two dimensions (i.e., attachment
anxiety and attachment avoidance) was the only model (Simpson et al., 1992) that fitted data
compared to categorical approaches. The dimensions of attachment anxiety and attachment
avoidance are understood by some researchers (see Fraley and Spieker, 2003) to be functionally
comparable to the child attachment categories of secure, anxious-resistant and avoidant as outlined
by Ainsworths original Strange Situation experiments (Ainsworth et al., 1978). Thus, using these
dimensions allows us to include studies using both parental and adult romantic partner attachment
measures. Moreover the use of anxious and avoidant orientations, which are theoretically
associated with hyperactivating and deactivating emotion regulation strategies respectively,
allowed us the opportunity to examine such associations within the literature.

Attachment and Emotion Regulation

A secure attachment style is hypothesised to develop when primary attachment figures respond
consistently to an individuals needs, by providing physical and emotional safety thus fulfilling an
individuals need for safety and comfort. This is dependent upon the attachment figures
availability, attunement to the child and the consistency of the attachment figures response.
Furthermore, a secure attachment style requires attachment figures to operate as a secure base
from which the child can explore the world (Bowlby, 1982). Primary attachment strategy involves
proximity seeking towards primary attachment figures. When the primary strategies of proximity
seeking fails, secondary strategies such as hyperactivating or deactivating strategies are employed.
An individual with a secure attachment style develops an internal working model that
internalises positive beliefs about the availability of others alongside beliefs about self-worth.
Internal working models are hypothesised to shape affect regulation by influencing the cognitive,
affective and behavioural responses individuals experience when distressed. For securely attached
individuals, beliefs about their primary attachment figures availability and ability to alleviate
distress are formed, in addition to beliefs that they are worthy of being cared for. This process
allows for the development of functional emotion regulation strategies through positive beliefs
about the self and positive beliefs about the abilities of others to relieve distress (Belsky, 2002;
Bowlby, 1973). Securely attached individuals, therefore, learn adaptive emotion regulation
strategies that allow them to manage distress in light of their previous positive experiences of
4
emotion regulation with primary caregivers. When primary attachment figures fail to respond to a
childs needs by responding in a way that does not relieve distress, or are unresponsive or absent,
an insecure or disorganised attachment style (Main, 1990) is developed. An individual with an
insecure attachment style develops a negative internal working model regarding themselves or
others. Individuals then develop strategies to regulate emotion and minimise threat towards
themselves according to the negative experiences they have had with their primary care givers.
Attachment styles are thus proposed to consist of internal working models influencing
emotion regulation and behaviour (Bretherton and Munholland, 1999). Avoidantly attached
individuals develop negative working models of others and anxiously attached individuals develop
negative working models of the self. Negative working models of the self, concern beliefs about
the acceptability and worth of the self, thus anxiously attached individuals are concerned they will
be rejected. Secondary hyperactivating strategies are thus employed which involve demanding
proximity and care from attachment figures (Mikulincer et al., 2003) and hyper-vigilance towards
attachment cues, the use of worry, and rumination (Cassidy, 1994). These strategies can often
amplify distress because they are maladaptive emotion regulation strategies (Mikulincer and
Shaver, 2007).
Negative internal working models regarding others are negative beliefs about the presence
and availability of attachment figures in addition to the expected response attachment figures will
give (for a review see Cassidy, 2000).
Avoidantly attached individuals find it difficult to be emotionally close to others and
struggle to depend on others. Secondary deactivating strategies are therefore employed such as
creating an emotional distance from others (Mikulincer et al., 2003) and the suppression of
distressing cognitions and memories (Shaver and Mikulincer, 2002).

Adult Attachment
Attachment styles are constructs developed in social psychology pertaining to Hazan and Shavers
(1987) proposition that adult romantic relationships show classifications similar to the attachment
classifications infants display.
The working models developed during childhood are thus hypothesised to influence
thoughts, feelings and behaviours in adults. As with primary caregivers, partners in adult
relationships are used as a safe base and sought to provide security and comfort (see Fraley and
5
Shaver, 2000). Caregivers can be experienced as a safe haven to return to when perceiving threat
or feeling anxious (Bowlby, 1988).
Insecure individuals thus are hypothesised to engage in hyperactivating or deactivating
emotion regulation strategies in attempts to meet attachment needs (Mikulincer et al., 2003).
Adults evaluations of childhood experiences and the influence of these childhood
experiences on functioning form attachment representations. Attachment representations also
determine the access or limitation of information regarding attachment (Main et al., 1985) thus
influencing attachment style.

Emotion Regulation
Emotion regulation is a process by which emotions are modulated both consciously and
unconsciously (Bargh and Williams, 2007; Rottenberg and Gross, 2003). The degree or type of
emotional experience is modified by individuals when they employ emotion regulation strategies
(Diamond and Aspinwall, 2003). Emotion regulation theory has expanded from stress and coping
research (Billings and Moos, 1981; Folkman and Lazarus, 1986) and early cognitive behavioural
theory (Beck, 1976).
There are often two types of emotion regulation specified in literature, adaptive and
maladaptive emotion regulation. Examples of adaptive emotion regulation include reappraisal of
positive or neutral interpretations of events and problem solving (Aldao et al., 2010), while
suppression and avoidance are examples of maladaptive emotion regulation (Gross, 1998; Hayes
et al., 2004). In Gross and Johns (2003) emotion regulation model, suppression, which is a
response-focussed strategy, has been noted to reduce the behavioural expression of negative
emotion, whilst still allowing the experience of negative emotion. Chronic suppression is
associated with an increased sensitivity to negative cognitions and depressive symptomatology,
while avoidance is associated with an increase in negative thoughts (Wenzlaff and Wegner, 2000),
another feature of depression. In contrast, reappraisal is described as an antecedent focussed
strategy that occurs early, thereby intervening before emotion is fully generated and allowing a
change in the course of the emotion regulation (Gross and John, 2003).
Research indicates the emotion regulation strategy of emotional disclosure, the sharing of
emotional experiences has a high negative correlation with suppression. Those with impaired
emotion regulation skills are therefore less likely to engage in emotional disclosure (Garrison et
6
al., 2012). Relatedly, emotion regulation strategies such as self-concealment, the concealing of
personal information, and emotional cut-off, distancing oneself from others (Wei et al., 2005a), are
considered maladaptive emotion regulation strategies. Experiences of depersonalisation, such as
self-splitting, as an emotion regulation strategy are also considered maladaptive (Lopez et al.,
2002).
Research has used coping measures to indicate adaptive emotion regulation, in addition to
concepts such as ego-resiliency, which is conceptualized as the ability to regulate negative feelings
in problem-solving and social contexts (Block, 1982; Kobak and Sceery, 1988). We used a broad
conceptualisation of emotion regulation for the purposes of this review because of the emerging
nature of this research.

Attachment, Dysfunctional Emotion Regulation and Symptoms of Depression

Attachment has been linked to depression in studies examining depressive symptomatology
(Gerlsma and Luteijn, 2000) and clinical depression (Hammen et al., 1995; Mickelson et al.,
1997). Symptoms of depression include low mood for most of the day, anhedonia, fatigue, feelings
of worthlessness or guilt, difficulty concentrating, indecision, recurrent thoughts of death,
insomnia or excessive sleep, and significant unintentional weight loss or weight gain (DSM IV
criteria for depression; American Psychiatric Association, 1994).
Anxious attachment orientations have also been associated with depression (Cole-Detke
and Kobak, 1996; Fonagy et al., 1996; Rosenstein and Horowitz, 1996; Zuroff and Fitzpatrick,
1995). Dismissing attachment has been found to be positively associated with depression (Patrick
et al., 1994). A North American study utilising a nationally representative sample reported on the
prevalence of major depressive episodes and the association of this with attachment styles
(Mickelson et al., 1997). Bivariate beta coefficients showed both anxious (b=0.34, p<.001) and
avoidant (b=0.42, p<.001) attachment orientations were associated with the prevalence of major
depressive episodes in a model using psychopathology to predict attachment styles. Similar
associations have been observed in Adolescent literature. Adolescents with anxious and avoidant
attachment orientations reported high levels of symptomatology in comparison to securely
attached adolescents (Cooper et al., 1998). Furthermore, adolescents with less effective emotion
regulation report more depressive symptomatology through the utilisation of emotion regulation
strategies such as rumination (Silk et al., 2003). Rumination can be defined as a process by which
7
the individual engages in repetitive thought regarding depressive symptoms, including thought
around the antecedents and consequences of such symptoms (Nolen-Hoeksema et al., 2008).
Depression is frequently considered a disorder that occurs due to a dysfunction of emotion
regulation strategies (Campbell-Sills and Barlow, 2007; Gross and Munoz, 1995; Mennin et al.,
2007). Research supports associations between major depressive disorder and emotion regulation
difficulties (Nolen-Hoeksema et al., 2008; Rottenberg et al., 2005). Cognitive emotion regulation
strategies, which are conceptualised as the cognitive coping strategies for the management of
emotions (Garnefski et al., 2001) have been significantly associated with depressive symptoms
(Garnefski and Kraaij, 2006). Rumination in particular has been examined as an emotion
regulation strategy that can lengthen episodes of depression (Nolen-Hoeksema et al., 2008) and
maintain negative affect (Nolen-Hoeksema, 2000; Roberts et al., 1998) and this process has been
viewed as a component of a broader thinking style for coping with inner events that leads to
psychological disorder (Wells, 2009; Wells and Matthews, 1994, 1996). However, other models
suggest depression is due to dysfunctional schemas about the self (Beck et al., 1979), which may
have their origins in early experiences of attachment. It is important as a starting point to assess the
nature of evidence linking attachment styles to maladaptive emotion regulation and depression
symptoms.

Method
A systematic search of EMBASE and PsycINFO was conducted using the initial search terms
attachment, emotion regulation and depression. A further search was conducted using the search
terms attachment, emotion regulation or affect regulation and depression, following which no
further relevant papers were identified. To ensure a thorough exploration of this literature, the
search was extended again by using the search terms attachment, suppression or disclosure or
rumination and depression, revealing four further papers. The search was initially conducted
across the lifespan. However, when only three relevant papers using child samples were extracted,
the inclusion criteria for the review was then revised to include adolescent and adult samples only
to maintain homogeneity across the reviewed studies.

8
Definition of Terms
As there is limited literature on specific types of emotion regulation, attachment and depressive
symptoms, and because of the emerging nature of the research, a broad definition of emotion
regulation was used including cognitive and emotion regulation strategies and coping mechanisms.
We therefore included a range of studies that were generated by emotion regulation or affect
regulation and the remaining search terms and did not exclude studies based on the type of
emotion regulation strategy utilised in studies. Given the range of self-report measures utilised in
attachment research, studies were included that used a range of attachment measures. Depressive
symptoms were informed by the symptoms described in DSM IV criteria for depression
(American Psychiatric Association, 1994).

Inclusion Criteria
The studies that were included met the following criteria: empirical study published in a peer
reviewed journal, from 1980 up to and including the year 2013, written in English, reporting on
attachment, emotion regulation and depression. Studies reporting on adolescent and adult samples
were included, which ranged from 8 to 43 years in the studies reviewed.
A study utilising a sample of participants ranging from 8 to 14 years was included in this
review, we have classified this as a study utilising an adolescent sample, as the sample in this
study was predominantly adolescent (Brenning et al., 2012).
The inclusion of adolescent studies in the review meant that a range of attachment
measures were used, particularly as measures of attachment can examine either parental
attachment or adult romantic partner attachment. Although this introduces potential developmental
differences, research supports the concept of continuity in attachment theory, whereby parental
attachment influences romantic partner attachment in adulthood (Pascuzzo et al., 2013).
Furthermore, the limited number of studies examining target variables as a result of the emerging
nature of this research meant that we were unable to exclude adolescent studies for the potential
developmental variation this may introduce. Instead, we have chosen to examine adolescent
studies and adult studies separately in order to collate associations found in papers with either
adolescent or adult samples. This has allowed us to examine both groups and to comment on
similarities across adult and adolescent samples.

9
Exclusion Criteria
Studies were excluded if they did not measure the constructs of attachment, emotion regulation or
coping, and depressive symptoms. All studies in this review measured these constructs in
participants, except for one adolescent study, which did not measure adolescent attachment but it
measured maternal attachment (Kobak and Ferenz-Gillies, 1995). Studies in languages other than
English were excluded alongside book chapters and dissertation abstracts. Three studies using
child samples were also excluded in order to maintain the homogeneity across the studies
reviewed. Although more recently, mindfulness has been described as an emotion regulation
concept we chose to exclude papers on this due to the debate over the exact components of
mindfulness (Baer et al., 2004).

Study Selection and Quality Rating

Overall 19 eligible papers were identified from 2087 papers (see Table A.1 for a summary of the
reviewed papers; see Figure A.2 for a schematic of the selection process). The relevance of 121
papers was assessed in a full text review but the majority were excluded because they did not
measure all constructs but a further five studies emerged from the inspection of reference lists
from those papers, resulting in 22 papers of which three were excluded because they used child
samples only.
The 19 papers identified were evaluated for methodological quality using the Checklist for
Measuring Quality (Downs and Black, 1998), by the first author (S.M.) and a peer, who was
independent of the research team and University. The Checklist for Measuring Quality has been
reported to have high internal consistency, good test-retest reliability and inter-rater reliability
(Downs and Black, 1998). As this quality assessment tool is applicable across a range of
methodological designs, including longitudinal and cross-sectional designs, it was deemed as the
most appropriate rating scale for this review.

[Insert Table A.1 and Figure A.2 about here]

The Checklist for Measuring Quality has 27 questions that give a maximum score of 32.
Questions cover the following areas: study reporting, internal validity, external validity, bias,
confounding, and power. Items such as Clear description of hypothesis/aims, Sample
10
representativeness of population, with responses or either yes or no, or scores of 1, 0 or unable to
determine (also scored as 0). As some items are related to randomisation and blinding, which
usually are only applicable to randomised controlled trial designs, studies using this methodology
will obtain scores for these items. Although randomised control trial (RCT) methodology is
considered a more valid design, the early nature of research in this area has meant that studies
using cross-sectional designs have been included in this review and their findings are not less
valid. However, there is an inherent bias towards RCT designs and other methodologies will
obtain lower scores on the quality rating tool for this reason. Consequently, we have provided a
detailed critical evaluation of the studies whilst using the checklist scores as a guide. Furthermore,
conclusions in this developing field of research are tentative and require more robust research
design to confirm relationships outlined in this paper.
Studies scoring greater than or equal to 20 are considered good, scores of between 15-19
are considered fair and studies scoring 14 or below are considered poor. Following initial rating by
the two raters, the intra-class correlation coefficient for the inter-rater reliability for the two sets of
ratings using the checklist was optimal at 0.89. Any discrepancies between ratings were resolved
following discussion and further examination of the papers by the two quality raters. The
following information was extracted from the selected papers: details about the sample, the size,
clinical/non-clinical status, age range; country the study was conducted in, the design of the study,
questionnaires used, main findings and quality rating using the checklist for measuring quality
(Downs and Black, 1998).

Results

Overview of Studies
The majority of the selected papers was North American (n=15), with a single study originating
from the Middle East, another study from Belgium, one from Germany and another from the
Netherlands. Sample sizes ranged from 53 to 2011 participants and their ages ranged from 8 to 43
years.

11
 Although studies using child samples were excluded, a study utilising a sample of
participants ranging from 8 to 14 years was included in this review, we have classified this as a
study utilising an adolescent sample (Brenning et al., 2012).
All studies used a quantitative design (n=19). Twelve studies reported on student samples
and seven reported on adolescents, most studies used normative samples (n=14). Most studies in
this review used established measures of depressive symptomatology, such as the Beck Depression
Inventory (BDI; Beck et al., 1961), the Depression Anxiety and Stress Scales (DASS; Lovibond
and Lovibond, 1995), the Inventory of Depression and Anxiety Symptoms (IDAS, Watson et al.,
2007), and the Center for Epidemiologic StudiesDepression Scale (CES-D; Radloff, 1977). Two
studies used more descriptive measures of mood, such as the Positive and Negative Affect Scale
(PANAS; Watson, et al., 1988) and the Depression Adjective Checklist, (DACL; Lubin, 1965).
Studies using adolescent samples used a range of age-appropriate measures to capture depressive
symptomatology including the Dimensions of Depression Profile for Adolescents (Harter et al.,
1987). One study used an idiosyncratic scale to measure mood following an experimental task.
A range of emotion regulation strategies was captured by various measures. Two studies
utilised the Distress Disclosure Index (DDI, Kahn and Hessling, 2001) measuring self-disclosure,
whilst four studies used coping measure of Problem-Focused Style of Coping measure which
captures suppressive and reactive style coping (PF-SOC; Heppner et al., 1995). Measures such as
the Negative Mood Regulation Scale (Catanzaro and Mearns, 1990), Emotion Regulation
Checklist (ERC; Shields and Cicchetti, 1997), the Child and Adolescent Problem Solving
Inventory (Gamble, 1993) and the Emotion Regulation Inventory (Roth et al., 2009) were also
used. The Differentiation of Self Inventory (DSI; Skowron and Friedlander, 1998) was used to
measure emotional cut-off and emotional reactivity and the Splitting Index (SI; Gould et al., 1996)
was used to measure self-splitting and self-concealment. The Taylor Manifest Anxiety Scale
Marlowe-Crowne Social Desirability Scale (Weinberger et al., 1979) was used in one study to
measure the willingness or ability to repress negative memories.
The studies included in this review have been grouped according to whether the sample
utilised was adult or adolescent. This allows for the collation of findings within these groups,
which may differ due to the variation in the use of parental or romantic partner attachment, and
developmental differences; within group homogeneity is therefore maintained.

12
Adolescent Studies of Associations between Attachment, Emotion Regulation and Depressive
Symptoms
Seven studies were identified that have examined relationships between the target variables in
adolescent samples (Brenning et al., 2012; Cooper et al., 1998; Kobak and Ferenz-Gillies, 1995;
Kullik and Petermann, 2013; Ruijten et al., 2011; Torquati and Vazsonyi, 1999; Zaremba and
Keiley, 2011).

Quality Ratings
Quality ratings for the seven adolescent studies, ranged from 12-17. Coopers and colleagues
(1998) study was scored as fair, whist the remaining studies were within the poor range
(Brenning et al., 2012; Kobak and Ferenz-Gillies, 1995; Kullik and Petermann, 2013; Ruijten et
al., 2011; Torquati and Vazsonyi, 1999; Zaremba and Keiley, 2011). Studies rated poor were
limited in the internal validity, external validity and reporting sections of the checklist for
measuring quality (Downs and Black, 1998).

Associations Between Target Variables

Six of the studies demonstrated significant positive associations between insecure attachment, poor
emotion regulation and depressive symptoms (Brenning et al., 2012; Cooper et al., 1998; Kullik
and Petermann, 2013; Ruijten et al., 2011; Torquati and Vazsonyi, 1999; Zaremba and Keiley,
2011). One study did not report on these associations (Kobak and Ferenz-Gillies, 1995), because
only maternal attachment was assessed and adolescent attachment was not directly measured.
Of the six studies, two studies reported on regression coefficients, b ranged from .06-.65
(Cooper et al., 1998; Torquati and Vazsonyi, 1999), showing positive associations between
attachment, emotion regulation and depressive symptoms that ranged widely in their magnitude.
One paper reported on two studies (Brenning et al., 2012). Study 1 in Brenning et al. (2012)
showed very weak to strong correlations between target variables (ranging from r=.19-.45, p <
.001); however, structural equation model analyses found emotion regulation did not mediate
attachment and depressive symptoms. Study 2 also reported on very weak to strong correlations
between target variables ranging from (r= .19-.53, p < .01) but reported an adequately fitting
structural equation model depicting partial mediation (df =113, SBS2 = 246.44, RMSEA = .05,
CFI = 0.99) showing emotion regulation partially mediated attachment and depressive symptoms.
13
 Two studies (Kullik and Petermann, 2013; Ruijten et al., 2011) showed positive
correlations between parental attachment, emotion regulation strategies and depressive symptoms
ranging from weak to moderate (r= .23-.47, p < .05). Both studies utilised bootstrapping to
examine mediation and found that parental attachment and depressive symptomatology
relationship was partially mediated by a) rumination (Ruijten et al., 2011) and b) both fully and
partially mediated by dysfunctional emotion regulation depending on the gender of participants
(Kullik and Petermann, 2013). Kullik and Petermann (2013) reported a mediator effect of internal
dysfunctional emotion regulation (i.e. I harm or punish myself in some way in order to regulate
emotions) for girls (4.71 b 13.91) in the relationship between attachment to parents and
depressive symptoms. In contrast they found partial mediation via internal-dysfunctional (3.04 b
12.71) and external dysfunctional (i.e. I take my feelings out of other people verbally to
regulate emotions) for boys when examining attachment to parents and depressive symptoms
(Kullik and Petermann, 2013).
The remaining study (Zaremba and Keiley, 2011) found a strong positive correlation
between maladaptive emotion regulation and internalising symptoms (r= .58, p<.001) and strong
negative intercorrelations for attachment security and maladaptive emotion regulation strategies
(r= -.42, p<.001), and attachment security and internalising behaviours (r= -.63, p<.001).
However, similar to Study 1 by Brenning and colleagues (2012), the final mediation model testing
emotion regulation as a mediator, did not fit (Zaremba and Keiley, 2011). This small sample of
adolescent studies thus show associations of varying strength between target variables and provide
some but inconsistent evidence for emotion regulation as a mediator. Specifically, two out of three
mediator studies found that an emotion-regulation model did not fit the data.

Analytic Strategy
An important factor applicable across all of the studies testing mediation is the choice of statistical
analysis used to test mediation. For example, regression analyses (Baron and Kenny, 1986), which
are commonly used to test mediation, have been found to have the lowest statistical power when
compared to other methods (MacKinnon et al., 2002). The Sobel test has been recommended as
powerful, but with problematic standard errors occurring through the use of specific statistical
programmes (MacKinnon et al., 2002). Alternatively, bootstrapping techniques (Efron and
Tibshirani, 1993; Preacher and Hayes, 2004) have been recommended for mediation analyses
14
(Preacher and Hayes, 2008; Shrout and Bolger, 2002). Furthermore, more recent recommendations
for mediation analyses outline the importance of computer-intensive resampling methods such as
Structural Equation Modelling within which bootstrapping is also often used (MacKinnon et al.,
2007). One study (Zaremba and Keiley, 2011) used mediation path analysis and another used
Structural Equation Modelling (Brenning et al., 2012), which are deemed powerful methods to test
mediation, particularly as these methods utilise bootstrapping and can measure indirect effects.
Two studies utilised recommended bootstrapping analyses (Kullik and Petermann, 2013; Ruijten
et al., 2011). The three remaining studies utilised less powerful techniques such as regression
analyses.

Findings Concerning Hyperactivating and Deactivating Strategies in Adolescent Samples

Findings concerning hyperactivating strategies have been reported where there are associations
between anxiously attached individuals and the employment of hyperactivating strategies and
depressive symptoms, for example, emotional reactivity. Findings concerning deactivating
strategies have been reported when there are associations between avoidantly attached participants
and the employment of deactivating strategies such as suppression, and depressive
symptomatology.
Two studies presented evidence for hyperactivating strategies. Anxiously attached
participants reported depressive symptoms (b=0.23) and were more likely to seek support (b=0.35;
Torquati and Vazsonyi, 1999). Emotion dysregulation was also presented as a mediator for
avoidant attachment and depressive symptoms in this study, which supports deactivating strategies
(Brenning et al., 2012).
Anxiously attached participants were reported to utilise maladaptive emotion regulation
through risky behaviours such as substance misuse and sexualised behaviour (b ranged from .06-
.65, Cooper et al., 1998), which may be conceptualised as a deactivating strategy, as such activity
can be used to avoid the experience of negative emotions. One study (Study 2, Brenning et al.,
2012) reported the use of suppression, a deactivating strategy that was a mediator for avoidant
attachment and depression, whilst two studies failed to provide evidence for hyperactivating and
deactivating strategies using mediation analyses (Study 1, Brenning et al., 2012; Zaremba and
Keiley, 2011). Brenning and colleagues (2012) attributed discrepancies in findings in studies 1 and
2 as due to the measurement of negative emotions in general in Study 1, rather than the specific
15
emotion regulation of sadness, which was measured in Study 2. They also noted that Study 2 had
a larger sample (N=746) with a wider age range (8-14 years) and assessed attachment with both
parents, whereas attachment to mothers was examined in Study 1, which may have led to
discrepancies. A further two studies (Kullik and Petermann, 2013; Ruijten et al., 2011) did not use
an attachment tool utilising anxious and avoidant attachment dimensions; results cannot therefore
be used as evidence for hyperactivating and deactivating strategies in anxious and avoidantly
attached individuals.

Attachment, Emotion Regulation and Depressive Symptoms: Studies Using Adult Samples
Twelve studies were identified that have examined relationships between the target variables in
adult samples (Burnette et al., 2009; Garrison et al., 2012; Kobak and Sceery, 1988; Lopez et al.,
2001, 2002; Marganska et al., 2013; McCarthy et al., 2006; Mikulincer and Orbach, 1995; Wei et
al., 2003, 2005a, 2005b 2006).

Quality Ratings
Quality ratings for adult studies ranged from poor to good with scores of 11-20 (Downs and Black,
1998). Five studies scored within the poor rating range with scores of 14 and below (Burnette et
al., 2009; Kobak and Sceery, 1988; Lopez et al., 2001; Marganska et al., 2013; Mikulincer and
Orbach, 1995). Six studies were given scores that were within the fair range of 15-19 (Garrison et
al., 2012; Lopez et al., 2002; McCarthy et al., 2006; Wei et al., 2003, 2005a, 2006). Finally, one
study (Wei et al., 2005b) obtained a high score of 20, indicative of a high quality methodological
study. All studies rated fair and good supported emotion regulation as a mediator of attachment
and symptoms of depression.

Associations Between Target Variables

Nine studies demonstrated significant positive associations between insecure attachment, poor
emotion regulation and depressive symptoms. The magnitude of the correlations ranged from r=
.08-.67, (p<.05), which indicated very weak to strong associations (Burnette et al., 2009; Garrison
et al., 2012; Lopez et al., 2001, 2002; Marganska et al., 2013; Wei et al., 2003, 2005a, 2005b,
2006).

16
 Four of these studies reported correlations ranging from moderate to strong, r= .30-.57,
p<.01 (Lopez et al., 2001, 2002; Wei et al., 2003, 2005a), whilst five studies showed a range of
correlations, including weak associations, r= .08-.56, p<.05 (Burnette et al., 2009; Garrison et al.,
2012; Marganska et al., 2013; Wei et al., 2005b, 2006). Studies thus provided evidence for
associations between target variables but demonstrated that these associations varied in strength
across studies.
One study (McCarthy et al., 2006) noted a similar link in the association of attachment
with participants expectancies of their abilities to regulate negative emotions, and stress
symptoms or emotions. Correlations ranged from r= .30-.36, p<.01. As emotion regulation was
not directly measured, we cannot assume this study provides evidence for the target variables
relevant for the review.

Findings Concerning Hyperactivating Strategies in Adult Samples

Seven studies demonstrated associations between anxious attachment, hyperactivating strategies
and depressive symptoms (Garrison et al., 2012; Lopez et al., 2001, 2002; Kobak and Sceery,
1988; Mikulincer and Orbach, 1995; Wei et al., 2005a, 2006). Correlations reported in these
studies ranged from r= .15 to .57 (p<.05) demonstrating a range of weak to strong correlations,
showing variation in target associations across studies.
Three studies (Garrison et al., 2012; Kobak and Sceery 1988; Mikulincer and Orbach,
1995) reported on the use of hyperactivating strategies in participants that were anxiously attached,
but did not test mediation for target variables. Hyperactivating strategies reported in those with
anxious attachment in these studies were: daily disclosure, when emotions are intensely
experienced, which is consistent with the strategy of seeking out others to regulate distress (b =
.12, p<.05; Garrison et al., 2012). Kobak and Sceery (1988) reported those that were anxiously
attached rated low on ego-resiliency, using a peer Q-sort emotion regulation measure, and reported
more distress than those in dismissing and secure attachment groups. The examination of anxious
attachment in Mikulincer and Orbachs (1995) experimental study was also consistent with
hyperactivating strategies, as anxious attachment was associated with low defensiveness and a
difficulty in repressing negative affect and cognition. This study, however, assessed transient
mood rather than depressive symptoms, thus has limited applicability to the associations we are
examining.
17
 Of these seven studies, four studies tested the role of hyperactivating emotion regulation
strategies using mediation analyses and found hyperactivating strategies to be mediators (Lopez et
al., 2001, 2002; Wei et al., 2005a, 2006).
A mediator for anxious attachment, negative mood and interpersonal problems, was
emotional reactivity (df = 93, RMSEA = .05, CFI=.98; Wei et al., 2005a). Reactive coping
(adjusted R2= .102, p<.01), which was a strategy aimed at seeking proximity to others, was also
a mediator for anxious attachment (adjusted R2= .226, p<.001) and distress (Lopez et al., 2001).
Attachment dimensions were entered into a regression model predicting distress (step1), reactive
and suppressive coping were then entered (step 2) (Lopez et al., 2001). At Step 1 anxious and
avoidant attachment scores combined accounted for 23% of the variance in distress (adjusted R2=
.226, p<.001), however avoidant attachment did not significantly predict distress, whilst anxious
attachment was a significant and positive predictor of distress (t(52) = 3.63, p < .001). At Step 2
the entry of the coping styles showed an enhancement of the prediction of distress scores and
explained a further 10% of the variance (adjusted R2= .102, p<.01) (Lopez et al., 2001).
Relatedly, Lopez et al. (2002) found the concealment of negative intimate information, i.e.
non-disclosure (adjusted R2= .02, p < .05), and self-splitting, (the experience of
depersonalisation and self-fragmentation; adjusted R2= .15, p <.001) and anxious attachment
(adjusted R2= .12, p < .001) were predictors in a regression model with the outcome of distress.
However, this study (Lopez et al., 2002) did not test mediation. Weis and colleagues (2006)
demonstrated that the relationships of anxious and avoidant attachment and depression are
mediated partially by the mediator of ineffective coping measured by suppressive and reactive
coping (scaled c2 = (df = 186) 240.58, p < .001, CFI = 1.00, RMSEA = .03) which provides
evidence for the both hyperactivating and deactivating strategies. Weis and colleagues study
(2003) found that perceived coping (which included reactive and suppressive coping) acted as a
mediator for both anxious and avoidant attachment and psychological distress, using a measure
that incorporated depressive symptoms (c2 = (df = 72) 241.56, p<.001, RMSEA = .07, CFI = .94).
This may provide evidence for both hyperactivating and deactivating strategies as both attachment
orientations and distress are mediated by perceived coping.

18
Findings Concerning Deactivating Strategies in Adult Samples
Four studies showed deactivating strategies acted as mediators in the relationship between
avoidant attachment and depressive symptoms (Marganska et al., 2013; Wei et al., 2005a, 2005b,
2006), with correlations ranging from weak to strong (r= .15-.67, p<.01). Avoidant attachment,
negative mood and interpersonal problems was mediated by deactivating strategies such as
emotional cut-off (Wei et al., 2005a, df = 93, RMSEA = .05, CFI=.98). The tendency to disclose
distressing information, which measures how comfortable the participant is with disclosing their
distress to others, was found to be a mediator for avoidant attachment and depression
longitudinally (c2 = (df = 156) 265.60, p <.01, CFI =.98; RMSEA = .05; Wei, 2005b).
Furthermore, the perceived inability to manage emotion responses, which is associated with the
experience of negative thoughts such as rumination fully mediated the relationship between
avoidant attachment and depression symptoms (b= .25, p<.05; Marganska et al., 2013).
One study, which did not test mediation, found attachment avoidance was negatively
associated with adaptive emotion regulation strategies such as generalised disclosure and daily
disclosure (b= -.14 and -.04, p<.05, Garrison et al., 2012). Daily disclosure was in turn found to be
negatively associated with depressive symptoms (Garrison et al., 2012; b= -.19, p<.05). Multi-
level models were used to analyse the data, which included the intensity of the emotional event as
a variable in the model and depressive symptoms were used as a predictor. These results are
coherent with deactivating strategies (Garrison et al., 2012). In contrast, one study using
regression analyses to test mediation (Lopez et al., 2002) failed to support coping style as a
mediator of avoidant attachment and depressive symptomatology. Another study examined
anxious and avoidant attachment, with rumination as a mediator to depressive symptoms, however
the mediation model used examined paths leading to the dependent variable of forgiveness. The
results for this study this cannot be used as evidence to support hyperactivating or deactivating
strategies as depressive symptoms are used as mediators themselves within this structural equation
model (Burnette et al., 2009).
Studies therefore provided consistent evidence for the associations between attachment,
depressive symptoms and hyperactivating strategies that were variable in strength. The mediation
of hyperactivating strategies with attachment and depressive symptoms was also consistently
found. Conversely, contradictory results were found amongst studies examining deactivating

19
strategies as results showed associations ranging in strength; with only some studies providing
evidence for mediation whilst others failed to do so.

Discussion
The aims of this paper were to systematically review literature on the relationships between
attachment, emotion regulation and depressive symptomatology and to explore emotion regulation
as mediator in the relationship between insecure attachment and depressive symptoms. The
relationships between specific dimensions of insecure attachment and types of emotion regulation
on depressive symptomatology were also explored.
Adolescent studies suggest variable and unreliable evidence for associations between
attachment, emotion regulation and depressive symptomatology. This includes contradictory
evidence of emotion regulation as a partial mediator for target variables in adolescent samples.
Evidence supporting the use of hyperactivating and deactivating strategies amongst anxious and
avoidantly attached adolescents respectively has been mixed. Associations between target
variables demonstrated in adult studies have also been variable, and unreliable, which may be due
to limitations in the measures and sampling methods utilised in the adult studies. Conversely,
evidence that emotion regulation is a mediator for target variables amongst methodologically
robust studies was strong. Hyperactivating strategies, in particular, have been consistently noted
as mediators for anxious attachment and depressive symptomatology, whereas evidence for
deactivating strategies as mediators between avoidant attachment and depressive symptoms has
been mixed.
The limited number of adolescent studies (n=7), of which a majority were rated as poor
(n=6), demonstrated a consistent association between attachment, emotion regulation, and
depressive symptoms, however, the strength of these relationships varied considerably, from very
weak to strong. Evidence for emotion regulation as a mediator of attachment and depressive
symptomatology was variable and unreliable, with studies both demonstrating mediation
(Brenning et al., 2012; Cooper et al., 1998; Kullik and Petermann, 2013; Ruijten et al. 2011;
Torquati and Vazsonyi, 1999) and failing to provide evidence of mediation (Brenning et al., 2012;
Zaremba and Keiley, 2011). Similarly, evidence for the use of hyperactivating strategies in
anxiously attached samples and evidence for avoidantly attached samples utilising deactivating
strategies was presented (Brenning et al., 2012; Cooper et al., 1998; Torquati and Vazsonyi, 1999).

20
However, two studies failed to confirm the use of hyperactiving and deactivating strategies
(Brenning et al., 2012; Zaremba and Keiley 2011).

Limitations of Adolescent Studies

Issues of conceptual singularity and statistical multicollinearity are applicable to all studies
included in this review.
Singularity occurs when two or more theoretical constructs are conceptually equivalent, in
the example of the studies reviewed, this conceptual equivalence may be between emotion
regulation and symptoms of depression (Gross and Thompson, 2007) or attachment and emotion
regulation. For example, emotion regulation strategies such as emotional cut-off may reflect
symptoms of depression and avoidant attachment may be a simply a marker of both.
Multicollinearity occurs when there is a high statistical correlation between measures (Hair, et al.,
2006). Problems with singularity and multicollinearity could mean studies detected associations
that were a result of conceptual overlap or high correlation between measures, rather than genuine
associations between differing constructs. Furthermore, the degree of singularity and collinearity
may have varied across studies as various measures were used, this in turn could affect the
reliability of findings. Two adolescent studies addressed the issue of multicollinearity and found
this did not interfere with results (Kullik and Petermann, 2013; Ruijten et al., 2011), problems with
conceptual singularity may still however have been problematic.

Study Design and the Generalizability of Samples

All adolescent studies employed a cross-sectional design, thus causality in the associations found
between target variables, cannot be inferred. All adolescent samples were drawn from normative
populations except for Zaremba and Keileys (2011) sample of adolescent males who had sexually
offended and reported anxiety, depression and somatic symptoms. This difference in samples
between all other studies and Zaremba and Keileys (2011) study, may explain the variation in
findings, although a continuum of processes has been suggested in common mental disorders,
which ranges from normative to clinical samples (Gibbs, 1996).
The sample sizes of the studies reviewed varied considerably from 62 participants
(Zaremba and Keiley, 2011) to 2011 participants (Cooper et al., 1998), which can affect
generalisability. Furthermore, generalisability was affected by the selection of participants from

21
female (Torquati and Vazsonyi, 1999) or male only samples that had sexually offended (Zaremba
and Keiley, 2011). The latter sample was also recruited from a larger intervention study at a
juvenile correlational facility, for which the inclusion criteria were not mentioned (Zaremba and
Keiley, 2011). Furthermore, treatment at this facility was reaching conclusion during the study
therefore the effects of this intervention were not accounted for and present a significant limitation
to the findings of this study. Selection bias may be present in Kobak and Ferenz-Gillies (1995)
study, which does not provide clarity regarding the selection of 42 families from the 54 families
that agreed to take part. The study least likely to be affected by selection bias and with the largest
sample size (Cooper et al., 1998) found that anxious attachment was associated with poor emotion
regulation and distress, when compared to secure and avoidant attachment. Sample limitations
also extend to ethnicity as the majority of participants in studies were Caucasian; additionally the
majority of studies were conducted in the United States which limits generalizability to the
countries and cultures within which the studies were conducted.

Limitations in the Measurement of Constructs

Four studies utilising adolescent samples used outcome measures specific to depression (Brenning
et al., 2012; Kullik and Petermann, 2013; Ruijten et al., 2011; Torquati and Vazsonyi, 1999). An
additional limitation to Zaremba and Kelieys (2011) findings, for the purposes of this review, is
their measurement of internalising symptoms, which includes the measurement of symptoms of
anxiety, thus presenting the possibility that associations of target variables with anxiety may have
confounded the results. This is a possibility because emotion regulation has been cited as a
mediator of insecure attachment and anxiety (Esbjrn et al., 2012). Furthermore, Cooper et al.
(1998) used the Brief Symptom Index (BSI; Derogatis and Melisaratos, 1983), which incorporates
depressive symptoms but also measures seven other types of symptoms including anxiety and
psychosomatic symptoms. Findings from Cooper et al. (1998) and Zaremba and Keiley (2011)
could thus be consequences of associations between target variables and outcomes such as anxiety,
or other types of symptomatology.
Reliable evidence has not been obtained for emotion regulation as a mediator, or the use of
hyperactivating and deactivating strategies, from studies utilising adolescent samples. This is a
consequence of the limitations these studies had in the areas of generalizability and sampling
through the use of male only or female only, largely Caucasian samples. Additionally, samples
22
varied considerably in size and one study utilised a clinical sample undergoing treatment (Zaremba
and Keiley, 2011). Choice of analytic strategy also limits the reliability of findings as all of the
studies did not utilise powerful analytic techniques such as bootstrapping or SEM. Additionally,
problems with conceptual singularity and statistical multicollinearity affect the reliability of the
studies.

Limitations of Adult Studies

The aforementioned problems with singularity, statistical multicollinearity and the relative power
of statistical analyses used to test mediation also are important factors to consider for adult studies.
Although issues of conceptual singularity were not addressed in any of the studies, an attempt to
examine multicollinearity was made in one study (Garrison et al., 2012), whilst remaining studies
did not address this issue. The correlations reported in this study (Garrison et al., 2012) suggested
that there was not enough overlap in constructs to lead to problems with multicollinearity.
One study (Lopez et al., 2001) used regression analyses, which has been deemed the least
powerful method of analysis to test mediation (MacKinnon et al., 2002). Three studies used a
recommended bootstrapping technique (Shrout and Bolger; 2002) and structural equation
modelling (SEM) to test mediation (Wei et al., 2005a, 2005b, 2006), whilst one study (Marganska
et al., 2013) utilised SEM and the sobel test for indirect effects. Although two studies (Burnette et
al., 2009; McCarthy et al., 2006) used a powerful method of analysis by using SEM, bootstrapping
was not used to test indirect effects of attachment through the mediator of emotion regulation as
recommended, thus providing relatively weaker analyses than studies that did use bootstrapping.
Another study used SEM and reported confidence intervals likely to have been obtained through
bootstrapping, although the use of this method is not explicitly reported in the paper (Wei et al.,
2003). Lastly, Mikulincer and Orbach (1995) used analysis of variance to compare groups and
Garrison et al. (2012) used multi-level modelling, but neither study directly tested mediation for
our target variables. Studies utilising more powerful bootstrapping methods thus found emotion
regulation mediated attachment and depressive symptoms (Wei et al., 2005a, 2005b, 2006).

23
Sample Limitations
Five studies (Burnette et al., 2009; Garrison et al., 2012; Lopez et al., 2001, 2002; Marganska et
al., 2013) reported disproportionate numbers of females included in their samples, which would
limit the findings of these studies to male populations. All adult studies used normative samples,
which limits the applicability to clinical populations. A range of sample sizes were utilised in
studies, which ranged from small to large (n= 53-515). There were two studies with small samples
of 53 (Kobak and Sceery, 1988) and 55 participants (Lopez et al., 2001); the findings for these
studies may be limited in terms of statistical power and the generalizability of results to the
populations samples were drawn from, as samples may not be representative. Five studies had
adequate samples of 120-229 participants (Burnette et al., 2009; Garrison et al., 2012; Lopez et al.,
2002; Mikulincer and Orbach, 1995; Wei et al., 2005a). An additional five studies utilised large
samples, which ranged from 284-515 participants (Marganska et al., 2013; McCarthy et al., 2006;
Wei et al., 2003, 2005b, 2006). Studies with large samples demonstrate greater generalizability
and statistical power; three of these studies provided evidence for emotion regulation as a mediator
(Marganska et al., 2013; Wei et al., 2005b, 2006).
All samples had a mean age of 18-20 years which is indicated a limited age range, a
consequence of sampling students. Findings are thus limited to students of a certain level of
academic achievement and to younger adults. Additionally, the samples included are limited in
ethnic diversity. Furthermore, as most of the studies were conducted in the United States, findings
have limited generalizability for other countries and cultures.

Measurement Limitations
Three studies (Lopez et al., 2001, 2002; McCarthy et al., 2006) used outcome measures that may
overlap with depressive symptoms such as distress and stress symptoms, but lack the specificity to
comprehensively assess target variables, as associations of emotion regulation with anxiety could
have confounded results (Esbjrn et al., 2012).
Amongst studies utilising adult samples, the studies that were rated to be methodologically
fair or good, used powerful statistical analysis and larger samples, found associations amongst
target variables that varied from weak to strong (Wei et al., 2005a, 2005b, 2006). Stronger
evidence for emotion regulation as a mediator was also found, with the consistent mediation of
hyperactivating strategies to anxious attachment and depressive symptoms (Lopez et al., 2001;
24
Wei et al., 2005a, 2006). The following hyperactivating strategies were found: emotional
reactivity, reactive coping, non-disclosure and self-splitting.
Evidence for the use of deactivating strategies such as emotional cut-off and reduced daily
disclosure, was also examined (Garrison et al., 2012; Marganska et al., 2013; Wei et al., 2005a,
2005b, 2006). Inconsistency in the evidence for the use of deactivating strategies was noted with
limitations to sampling, generalizability and strength of chosen analytic strategy. These studies
failed to find a mediating role for deactivating strategies, such as self-splitting, self-concealment
and suppressive coping (Lopez et al., 2001, 2002). It thus appears that there is stronger evidence
for use of hyperactivating strategies and variable evidence for use of deactivating strategies in
adults. This finding is supported by the aforementioned literature linking anxious attachment to
depression (Cole-Detke and Kobak, 1996; Fonagy et al., 1996; Rosenstein and Horowitz, 1996;
Zuroff and Fitzpatrick, 1995) and theories that depression is associated with the dysfunction of
emotion regulation (Campbell-Sills and Barlow, 2007; Gross and Munoz, 1995; Mennin et al.,
2007).
However, it is important to interpret these data within the context of unresolved issues of
singularity and multicollinearity, which means evidence for emotion regulation mediation and the
use of hyperactivating and deactivating strategies must be tentatively presented.

Limitations of the Review

Problems with singularity and multicollinearity make it is difficult to determine if the findings of
the studies were due to associations between constructs that are conceptually distinct or whether
results are a consequence of issues with singularity or high levels of intercorrelation between
measures. Furthermore, researchers have argued that multicollinearity in studies using structural
equation modelling can lead to inaccurate parameter estimates and large standard errors of such
estimates (Grapentine, 2000), which leads to a difficulty in reliably interpreting the results of
analyses of studies included in the review. However, emotion regulation and attachment have been
found to be related but distinct concepts.
Studies demonstrated these constructs had a shared variance ranging from 6-27% (Lopez,
2001, 2002; Wei et al., 2003). Nevertheless difficulties with singularity and multicollinearity may
still exist and require consideration when discussing findings from studies examining these factors.
25
 The psychometric measures used in all of the reviewed studies were predominantly self-
report measures.
Self-report measures are often subject to retrospective bias and could lead to participants reporting
socially desirable outcomes (Pervin, 1999).
Participants exhibiting repressive defensiveness are particularly vulnerable to this bias (Davis,
1987). This may have led to reporter bias across studies. Moreover, one recognised limitation
across the attachment literature is the use of various assessment measures that produce differing
categories and conceptualisations of insecure attachment styles. The Adult Attachment Interview
is acknowledged as the gold standard measure (George et al., 1986). However, as an interview-
based measure, it is lengthy to administer. Due to feasibility, quick to administer self-report
measures are often used in preference to the AAI, but the various insecure attachment categories
endorsed by attachment measures can make it difficult to amalgamate findings.
There was heterogeneity in the emotion regulation strategies reported in the studies
examined in this review. Conducting research on specific cognitive emotion regulation techniques
and coping would provide further insight into the specific mechanisms at play between the
constructs of attachment and depression.
The majority of the studies reviewed are from non-clinical samples; this highlights the
emerging nature of the research but it also limits the generalizability of the findings of this review.
Additionally, the majority of studies were conducted in North America (n=13). Studies examining
the emotion regulation as a mediator for insecure attachment and distress within clinically
depressed populations, in various countries, would strengthen our conclusions.
Three studies using child samples (Cohen et al., 1999; Kidwell et al., 2010; Moss et al.,
2011) were excluded in order to maintain homogeneity across reviewed studies, particularly as
literature suggests emotion regulation strategies can vary amongst children according to age.
According to Perlman and Pelphrey (2010), younger children use more ventral emotional areas
of the brain during emotion regulation, whereas older children utilise more dorsal cognitive areas
when regulating their emotions. Changes in emotion regulation in children could thus lead to
difficulties in the interpretation of the results of studies using child samples.
A final limitation relates to the quality appraisal of studies and the fact that the tool used
inflated the score for studies using RCTs. Hence, the score should be considered as a guide only.

26
Implications for Research
Further longitudinal studies using clinically depressed populations, using self-report and third
party measures for the constructs of attachment, emotion regulation and depression, and utilising
powerful statistical techniques, are required to strengthen conclusions. Additionally, discrepancies
in the findings of studies examining deactivating strategies could be elucidated through future
research utilising robust methodology. Furthermore, research is also required which examines
target variables in adolescent samples.
Gaining further understanding of which specific emotion regulation strategies emerge from
anxious or avoidant attachment, and how these pathways lead to depressive symptoms, will assist
clinicians in the development of specific emotion regulation interventions. Focussing interventions
on mediators between early attachment and psychological distress could pose an alternative to
more resource intensive interventions that may be aimed at correcting the effects of negative early
attachment experiences. In order to avoid problems of conceptual singularity and statistical
multicollinearity future studies could test more specific theoretically grounded pathways from
attachment to symptomatology that focus on underlying self-control processes. Specifically, the
metacognitive model of psychopathology (Wells, 2009; Wells and Matthews, 1994) assigns
explicit roles to styles of regulating cognition as causal factors in disorder. Problematic worry and
rumination-focused styles are thought to result from metacognitive beliefs, which could in part
have origins in attachment styles. Alternatively, the role of schemas (Beck et al., 1979) concerning
the self and others might further clarify the pathway by which individuals with anxious or avoidant
attachment develop symptomatology, whilst avoiding the methodological limitations that occur
when testing emotion regulation as a lone mediator.

Clinical Implications
The results indicate that hyperactivating strategies are used by adults presenting with depression
and anxious attachment. The manipulation of maladaptive hyperactivating strategies in anxiously
attached service users suffering from depression might facilitate a reduction in depressive
symptomatology. Implications from the findings of this review indicate that future research
should examine interventions aimed at changing hyperactivating strategies in anxiously attached
depressed samples.

27
 Furthermore, research has indicated that therapists knowledge of clients use of
hyperactivating and deactivating strategies can be used to inform and maintain therapeutic
relationships (Mallinckrodt, 2010). The Gratification, Relief, Anxiety, Frustration model (GRAF
model) outlines how therapists may regulate therapeutic distance in order to facilitate change in
clients that employ hyperactivating and deactivating strategies (Mallinckrodt, 2010). Since the
findings from this review indicate that in adult samples hyperactivating strategies mediate between
anxious attachment and depressive symptoms. Future studies could test the GRAF model for adult
service users suffering from depression and displaying anxious attachment, which in turn could
inform the development of the GRAF model (Mallinckrodt, 2010). Furthermore, evidence of the
role of rumination as a mediator to avoidant attachment and depressive symptoms calls for the use
of models such as the metacognitive model of depression which focus upon rumination (Wells,
2009).
There are more general assessment and measurement implications that have emerged form
our analysis pointing to the importance of developing measures of attachment, self-regulation and
symptoms that are more clearly defined and differentiated in order to examine relationships
between these constructs in the future.

Conclusion
Evidence of relationships between attachment, emotion regulation and symptoms of depression is
variable and unreliable for studies with adolescent samples. There has also been contradictory
evidence of emotion regulation as a partial mediator for target variables and for the use of
hyperactivating strategies and deactivating strategies amongst adolescent studies. Associations
demonstrated in adult studies have also been variable, and unreliable, which may be due to
limitations in the measures and sampling methods used. Conversely, evidence that emotion
regulation is a mediator for target variables amongst methodologically robust studies was strong.
This is consistent with evidence that emotion regulation is a mediator for attachment and anxiety
(Esbjrn et al., 2012). Hyperactivating strategies, in particular, have been consistently noted as
mediators for anxious attachment and depressive symptomatology, whereas evidence for
deactivating strategies as mediators between avoidant attachment and depressive symptoms has
been mixed.
28
References

Aldao, A., Nolen-Hoeksema, S., Schweizer, S., 2010. Emotion-regulation strategies across

psychopathology: A meta-analytic review. Clin. Psychol .Rev. 30, 217-237.

Ainsworth, M. D. S., Blehar, M. C., Waters, E., Wall, S., 1978. Patterns of attachment: A

psychological study of strange situation and at home. Hillsdale, NJ: Erlbaum.

American Psychiatric Association, 1994. Diagnostic and statistical manual of mental disorders.

(4th ed.). Washington, DC: Author.

Baer, R. A., Smith, G. T., Allen, K. B., 2004. Assessment of mindfulness by self-report: The

Kentucky Inventory of Mindfulness Skills. Assessment 11, 191206.

Bargh, J. A., Williams, L. E., 2007. On the nonconscious of emotion regulation. In: Gross, J. J.

(Ed.), Handbook of Emotion Regulation, Guilford Press, NY, pp. 429444.

Baron, M. R., Kenny, D. A., 1986. The moderatormediator variable distinction in social

psychological research: conceptual, strategic, and statistical considerations. J. Pers.Soc.

Psychol. 51, 11731182.

Bartholomew, K., Horowitz, L. M., 1991. Attachment styles among young adults: A test of a

four-category model. J. Pers. Soc. Psychol. 61, 226244.

Beck, A. T., 1976. Cognitive therapy and the emotional disorders. New York International

Universities Press, NY.

Beck, A. T., Rush, A. J., Shaw, B. F., Emery, G., 1979. Cognitive therapy of depression.

Guilford, NY.

Beck, A. T., Ward, C. H., Mendelson, M., Mock, J., Erbaugh, J., 1961. An inventory measuring

depression. Arch. Gen. Psychiat. 4, 561-571.

29
Belsky, J., 2002. Developmental origins of attachment styles. Attach. Hum. Dev. 4, 166170.

Besser, A., Priel, B., 2003. A multisource approach to self-critical vulnerability to depression: The

moderating role of attachment. J. Pers. 71, 515555.

Billings, A. G., Moos, R. H., 1981. The role of coping responses and social resources in

attenuating the stress of life events. J. Behav. Med. 4, 139157.

Block, J., 1982. Assimiliation, accommodation, and the dynamics of personality development.

Child Dev. 53, 281-295.

Bowlby, J., 1973. Attachment and loss: Vol. 2. Separation: Anxiety and anger. Penguin Books,

London.

Bowlby, J., 1980. Attachment and loss: Vol. 3. Loss: Sadness and depression. Basic Books, NY.

Bowlby, J., 1982. Attachment and loss: Vol. 1. Attachment (2nd Ed.). Basic Books, NY.

Bowlby, J., 1988. A secure base: Clinical applications of attachment theory. Routledge, London.

Brennan, K. A., Clark, C. L., Shaver, P. R., 1998. Self-report measurement of adult attachment:

An integrative overview. In: Simpson, J. A. and Rholes, W. S. (Eds.), Attachment theory

and close relationships. Guilford Press, NY, pp. 4676.

Brenning, K.M., Soenens, B., Braet, C., Bosmans, G., 2012. Attachment and depressive

symptoms in middle childhood and early adolescence: Testing the validity of the emotion

regulation model of attachment. Pers. Relationship. 19, 445-464.

Bretherton, I., Munholland, K. A., 1999. Internal working models in attachment relationships: A

construct revisited. In: Cassidy, J. and Shaver, P. R. (Eds.), Handbook of attachment

theory, research and clinical applications. Guilford Press, NY, pp. 89 114.

30
Burnette, J.L., Davis, D.E., Green, J.D., Worthington Jr, E.L., Bradfield, E., 2009. Insecure

attachment and depressive symptoms: The mediating role of rumination, empathy and

forgiveness. Pers. Indiv. Differ. 46, 276-280.

Campbell-Sills, L., Barlow, D. H., 2007. Incorporating emotion regulation into conceptualizations

and treatments of anxiety and mood disorders. In: Gross, J. J. (Ed.), Handbook of emotion

regulation. Guilford Press, NY, pp. 542559.

Carnelley, K. B., Pietromonaco, P. R., Jaffe, K., 1994. Depression, working models of others, and

relationship functioning. J. Pers. Soc. Psychol. 66, 127140.

Cassidy, J., Kobak, R., 1988. Avoidance and its relationship with other defensive processes. In:

Belsky, J. and Nezworski, T. (Eds.), Clinical implications of attachment. Hillsdale,

Erlbaum, NJ, pp. 300 - 323.

Cassidy, J., 1994. Emotion regulation: Influences of attachment relationships. Monogr. Soc. Res.

Child 59, 228-249.

Cassidy, J., 2000. Adult romantic attachments: A developmental perspective on individual

differences. Rev. Gen. Psychol. 4, 111-131.

Catanzaro, S. J., Mearns, J., 1990. Measuring generalized expectancies for negative mood

regulation: Initial scale development and implications. J. Pers. Assess. 54, 546-563.

Cohen, N.J., Muir, E., Lojkasek, M., Muir, R., Parker, C.J., Barwick, M., Brown, M., 1999.

Watch, wait and wonder: testing the effectiveness of a new approach to mother-infant

psychotherapy. Infant Ment. Health J. 20, 429-451.

Cole-Detke, H., Kobak, R., 1996. Attachment processes in eating disorder and depression. J.

Consult. Clin. Psych. 64, 282-290.

31
Collins, N. L., Ford, M. B., Guichard, A. C., Allard, L. M., 2006. Working models of attachment

and attribution processes in intimate relationships. Pers. Soc. Psychol. B. 32, 201219.

Cooper, M. L., Shaver, P. R., Collins, N. L., 1998. Attachment styles, emotion regulation, and

adjustment in adolescence. J. Pers. Soc. Psychol. 74, 1380-1397.

Davis, P. J., 1987. Repression and the inaccessibility of affective memories. J. Pers. Soc. Psychol.

53, 585-593.

Derogatis, L. R., Melisaratos, N., 1983. The Brief Symptom Inventory: An introductory report.

Psychol. Med. 13, 595-605.

Diamond, L. M., Aspinwall, L. G., 2003. Emotion regulation across the life span: An integrative

perspective emphasizing self-regulation, positive affect, and dyadic processes. Motiv.

Emotion 27, 125156.

Downs, S.H., Black, N., 1998. The feasibility of creating a checklist for the assessment of the

methodological quality both of randomized and non-randomized studies of health care

interventions. J. Epidemiol. Commun. H. 52, 377-384.

Efron, B., Tibshirani, R., 1993. An introduction to the bootstrap. Chapman & Hall/CRC, NY.

Esbjrn, B. H., Bender, P. K., Reinholdt-Dunne, M. L., Munck, L. A., Ollendick, T. H., 2012. The

development of anxiety disorders: considering the contributions of attachment and emotion

regulation. Clin. Child Fam. Psych. 15, 129-143.

Feeney, J., Noller, P., 1996. Adult attachment. Thousand Oaks, Sage, CA.

Folkman, S., Lazarus, R. S., 1986. Stress processes and depressive symptomatology. J. Abnorm.

Psychol. 95, 107113.

32
Fonagy, P., Leigh, T., Steele, M., Steele, H., Kennedy, R., Mattoon, G., Target, M., Gerber, A.,

1996. The relation of attachment status, psychiatric classifications, and response to

psychotherapy. J. Consult. Clin. Psych. 64, 23-31.

Fraley, R. C., Waller, N. G., 1998. Adult attachment patterns: A test of the typological model. In:

Simpson, J. A. and Rholes, W. S. (Eds.), Attachment theory and close relationships.

Guilford Press, NY, pp. 77114.

Fraley, R.C., Shaver, P.R., 2000. Adult Romantic Attachment: Theoretical Emerging

Controversies, and Unanswered Questions. Rev. Gen. Psychol. 4(2), 132-154.

Fraley, R.C., Spieker, S.J., 2003. Are infant attachment patterns continuously or categorically

districbuted? A taxometric analysis of strange situation behavior. Dev. Psychol. 39(3), 387-

404.

Garrison, A. M., Kahn, J. H., Sauer, E. M., Florczak, M. A., 2012. Disentangling the effects of

depression symptoms and adult attachment on emotional disclosure. J. Couns. Psychol.

59(2), 230-239.

Garnefski, N., Kraaij, V., 2006. Relationships between cognitive emotion regulation strategies

and depressive symptoms: A comparative study of five specific samples. Pers. Indiv.

Differ. 40(8), 1659-1669.

Garnefski, N., Kraaij, V., Spinhoven, P. H., 2001. Negative life events, cognitive emotion

regulation and depression. Pers. Indiv. Differ. 30, 13111327.

George, C., Kaplan, N., Main, M., 1986. Adult Attachment Interview. Unpublished.

Gerlsma, C., Luteijn, F., 2000. Attachment style in the context of clinical and health psychology: a

proposal for the assessment of valence, incongruence, and accessibility of attachment

representations in various working models. Brit. J. Med. Psychol. 73, 1534.

33
Gibbs, N. A., 1996. Nonclinical populations in research on obsessive-compulsive disorder: A

critical review. Clin. Psychol. Rev. 16, 8, 729-773.

Gould, J. R., Prentice, N. M., Ainslie, R. C., 1996. The splitting index: Construction of a scale

measuring the defense mechanism of splitting. J. Pers. Assess. 66, 414-430.

Grapentine, T., 2000. Path analysis vs. structural equation modeling. Marketing Research, 12(3),

1220.

Gross, J. J., 1998. The emerging field of emotion regulation: An integrative review. Rev. Gen.

Psychol. 2, 271299.

Gross, J. J., John, O. P., 2003. Individual differences in two emotion regulation processes:

implications for affect, relationships and well-being. J. Pers. Soc. Psychol. 85(2), 348-362.

Gross, J. J., Munoz, R. F., 1995. Emotion regulation and mental health. Clin. Psychol.-Sci. Pr 2,

151164.

Gross, J.J., Thompson, R.A., 2007. Emotion regulation: Conceptual foundations. In: Gross, J. J.

(Ed.), Handbook of emotion regulation, Guilford Press, NY, pp. 3-24.

Hair, J. F., Black, W.C., Babin, B. J., Anderson, R.E., Tatham, R. L., 2006. Multivariate Data

Analysis (6th Edition) Upper Saddle River, Prentice Hall, NJ.

Hammen, C. L., Burge, D., Daley, S. E., Davila, J., Paley, B., Rudolph, K. D., 1995. Interpersonal

attachment cognitions and prediction of symptomatic responses to interpersonal stress. J.

Abnorm. Psychol. 104, 436443.

Harter, S., Marold, P., Nowakowski, M., 1987. Manual for the dimensions of depression profile

for children and adolescents. University of Delfiver, Department of Psychology.

Hayes, S. C., Strosahl, K. D., Wilson, K. G., Bissett, R. T., Pistorello, J., Toarmino, D., Polusny,

M. A., Dykstra, T. A., Batten, S. V., Bergan, J., Stewart, S. H., Zvolensky, M. J., Eifert, G.

34
 H., Bond, F. W., Forsyth J. P., Karekla, M., McCurry, S. M., 2004. Measuring experiential

avoidance: A preliminary test of a working model. Psychol. Rec. 54, 553578.

Hazan, C., Shaver, P. R., 1987. Romantic love conceptualized as an attachment process.
J Pers Soc Psychol, 52, 511524.

Heppner, P. P., Cook, S. W., Wright, D. M., Johnson, W. C., Jr., 1995. Progress in resolving

problems: A problem-focused style of coping. J. Couns. Psychol. 42, 279293.

Kahn, J. H., Hessling, R. M., 2001. Measuring the tendency to conceal versus disclose

psychological distress. J. Soc. Clin Psychol. 20, 41-65.

Kidwell, S. L., Young, M. E., Hinkle, L. D., Ratliff, A. D., Marcum, M. E., Martin C. N., 2010.

Emotional competence and behavior problems: Differences across preschool assessment of

attachment classifications. Clin. Child Psychol. Psychiat. 15, 391-406.

Kobak, R., Sceery, A., 1988. Attachment in late adolescence: Working models, affect regulation,

and representations of self and others. Child Dev. 59, 135146.

Kobak, R., Ferenz-Gillies, R., 1995. Emotion regulation and depressive symptoms during

adolescence: a functionalist perspective. Dev. Psychopathol. 7, 183-192.

Kullik, A., Petermann, F., 2013. Attachment to parents and peers as a risk factor for adolescent

depressive disorders: the mediating role of emotion regulation. Child Psychiat. Hum. D.

44, 537-548.

Kurdek, L. A., 2002. On being insecure about the assessment of attachment styles. J. Soc. Pers.

Relations. 19(6), 811-834.

Lopez, F.G., Mauricio, A.M., Gormley, B., Simko, T., Berger, E., 2001. Adult attachment

orientations and college student distress: the mediating role of problem coping styles. J.

Couns. Dev. 79, 459-464.

35
Lopez, F. G., Mitchell, P., Gormley, B., 2002. Adult attachment and college student distress: Test

of a mediational model. J. Couns. Psychol. 49, 460467.

Lovibond, S. H., Lovibond, P. F., 1995. Manual for the Depression Anxiety and Stress Scales.

Psychological Foundation of Australia, Sydney.

Lubin, B., 1965. Adjective checklists for measurement of depression. Arch. Gen. Psychiat 12, 57

62.

MacKinnon, D.P., Fairchild, A.J., Fritz, M. S., 2007. Mediation Analysis. Annu. Rev. Psychol. 58,

593-614.

MacKinnon, D. P., Lockwood, C. M., Hoffman, J. M., West, S. G., Sheets, V., 2002. A

comparison of methods to test mediation and other intervening variable effects. Psychol.

Methods 7, 83104.

Main, M., Kaplan, N., Cassidy, J., 1985. Security in infancy, childhood and adulthood: A move to

the level of representations. Monogr. Soc. Res. Child. Dev.,50, 1-2.

Main, M., 1990. Cross-cultural studies of attachment organization: Recent studies, changing

methodologies, and the concept of conditional strategies. Hum. Dev. 33, 48-61.

Mallinckrodt, B., 2010. The psychotherapy relationship as attachment: Evidence and implications.

J. Pers. Soc. Relations. 27, 262-270.

Marganska, A., Gallagher, M., Miranda, R., 2013. Adult attachment, emotion dysregulation, and

symptoms of depression and generalized anxiety disorder. Am. J. Orthopsychiat 1, 131-

141.

McCarthy, C.J., Lambert, R.G., Moller, N.P., 2006. Preventive resources and emotion regulation

expectancies as mediators between attachment and college students stress outcomes. Int.

J. Stress Manage. 13(1), 1-22.

36
Mennin, D. S., Holoway, R. M., Fresco, D. M., Moore, M. T., Heimberg, R. G., 2007. Delineating

components of emotion and its dysregulation in anxiety and mood psychopathology.

Behav. Ther. 38, 284302.

Mickelson, K.D., Kessler, R.C., Shaver, P. R., 1997. Adult attachment in a nationally

representative sample. J. Pers. Soc. Psychol. 73, 1092-1106.

Mikulincer, M., Orbach, I., 1995. Attachment styles and repressive defensiveness: The

accessibility and architecture of affective memories. J. Pers. Soc. Psychol. 68(5), 917-925.

Mikulincer, M., Shaver, P. R., Pereg, D., 2003. Attachment Theory and Affect Regulation: The

Dynamics, Development, and Cognitive Consequences of Attachment-Related Strategies.

Motiv. Emotion 27(2), 77-102.

Mikulincer, M., Shaver, P. R., 2007. Attachment in adulthood: Structure, dynamics, and change.

The Guilford Press, NY.

Moss, E., Dubois-Comtois, K., Cyr, C., Tarabulsy, G.M., St-Laurent, D., Bernier, A., 2011.

Efficacy of a home-visiting intervention aimed at improving maternal sensitivity, child

attachment and behavioural outcomes for maltreated children: a randomized control trial.

Dev. Psychopathol. 23, 195-210.

Nolen-Hoeksema, S., 2000. The role of rumination in depressive disorders and mixed

anxiety/depressive symptoms. J. Abnorm. Psychol 109, 504-511.

Nolen-Hoeksema, S., Wisco, B. E., Lyubomirsky, S., 2008. Rethinking rumination. Perspect.

Psychol. Sci. 3, 400424.

Pascuzzo, K., Cyr, C., Moss, E., 2013. Longitudinal association between adolescent attachment,

adult romantic attachment, and emotion regulation strategies. Attach. Hum. Dev. 15(1), 83-

103.

37
Patrick, M., Hobson, R. P., Castle, D., Howard, R., Maughan, B., 1994. Personality disorder and

the mental representation of early social experience. Dev. Psychopathol. 6, 375-388.

Perlman, S. B., Pelphrey, K. A., 2010. Regulatory Brain Development: Balancing Emotion and

Cognition. Soc. Neurosc. 5(5-6), 533-542.

Pervin, L. A., 1999. Epilogue: Consistency and change in personality theory and research. In:

Pervin L. A. and John, O. P. (Eds.), Handbook of personality (2nd Ed.). Guilford Press,

NY, pp. 689704.

Preacher, K. J., Hayes, A. F., 2004. SPSS and SAS procedures for estimating indirect effects in

simple mediation models. Behav. Res. Meth. Instr. C. 36, 717-731.

Preacher, K. J., Hayes, A. F., 2008. Asymptotic and resampling strategies for assessing and

comparing indirect effects in multiple mediator models. Behav. Res. Methods 40, 879-891.

Radloff, L. S., 1977. The Center for Epidemiologic Studies Depression Scale (CES-D) for research

in the general population. Appl. Psych. Meas. 1, 385401.

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. J. Pers.

Soc. Psychol. 70, 310320.

Roberts, J. E., Gilboa, E., Gotlib, I. H., 1998. Ruminative response style and vulnerability to

episodes of dysphoria: Gender, neuroticism, and episode duration. Cognitive Ther. Res. 22,

401-423.

Roisman, G. I., Holland, A., Fortuna, K., Fraley, R. C., Clausell, E., Clarke, A., 2007. The Adult

Attachment Interview and self-reports of attachment style: An empirical rapprochement. J.

Pers. Soc. Psychol. 92(4), 678.

38
Rosenstein, D., S., Horowitz, H. A., 1996. Adolescent attachment and psychopathology. J.

Consult. Clin. Psych. 64, 244-253.

Roth, G., Assor, A., Niemiec, C. P., Ryan, R. M., Deci, E. L., 2009. The emotional and academic

consequences of parental conditional regard: Comparing conditional positive regard,

conditional negative regard, and autonomy support as parenting practices. Dev. Psychol.

45, 11191142.

Rottenberg, J., Gross, J. J., 2003. When emotion goes wrong: Realizing the promise of affective

science. Clin. Psychol.-Sci. Pr. 10, 227232.

Rottenberg, J., Gross, J. J., Gotlib, I. H., 2005. Emotion context insensitivity in major depressive

disorder. J. Abnorm. Psychol. 114, 627639.

Ruijten, T., Roelofs, J., Rood, L., 2011. The Mediating Role of Rumination in the Relation

Between Quality of Attachment Relations and Depressive Symptoms in Non-Clinical

Adolescents. J. Child Fam. Studies 20, 452-459.

Shaver, P. R., Mikulincer, M., 2002. Attachment-related psychodynamics. Attach. Hum. Dev. 4,

133161.

Silk, J. S., Steinberg, L., Morris, A. S., 2003. Adolescents Emotion Regulation in Daily Life:

Links for Depressive Symptoms and Problem Behaviour. Child Dev. 74(6), 1869-1880.

Simpson, J. A., Rholes, W. S., Nelligan, J. S., 1992. Support seeking and support giving within

couples in an anxiety-provoking situation: The role of attachment styles. J. Pers. Soc.

Psychol. 62(3), 434.

Shields, A., Cicchetti, D., 1997. Emotion regulation among school-age children: The development

and validation of a new criterion q-sort scale. Dev. Psychol. 33, 906-916.

39
Shrout, P. E., Bolger, N., 2002. Mediation in experimental and non-experimental studies: New

procedures and recommendations. Psychol. Methods 7, 422445.

Skowron, E. A., Friedlander, M. L., 1998. The Differentiation of Self Inventory: Development and

initial validation. J. Couns. Psychol. 45, 235-246.

Torquati, J.C., Vazsonyi, A. T., 1999. Attachment as an organizational construct for affect,

appraisals, and coping of late adolescent females. J. Youth Adolescence 28(5), 545-562.

Watson, D., Clark, L. A., Tellegen, A., 1988. Development and validation of brief measures of

positive and negative affect: The PANAS scales. J. Pers. Soc. Psychol. 54(6), 1063-1070.

Watson, D., OHara, M. W., Simms, L. J., Kotov, R., Chmielewski, M., McDade-Montez, E. A.,

Gamez, W., Stuart, S., 2007. Development and validation of the Inventory of Depression

and Anxiety Symptoms (IDAS). Psychol. Assessment 19, 253268.

Wei, M., Heppner, P. P., Mallinckrodt, B., 2003. Perceived coping as a mediator between

attachment and psychological distress: A structural equation modeling approach. J. Couns.

Psychol. 50, 438447.

Wei, M., Mallinckrodt, B., Russell, D. W., Abraham, T. W., 2004. Maladaptive perfectionism as a

mediator and moderator between attachment and negative mood. J. Couns. Psychol. 51,

201212.

Wei, M., Vogel, D.L., Ku, T-Y., Zakalik, R.A., 2005a. Adult attachment, affect regulation,

negative mood, and interpersonal problems: The mediating roles of emotional reactivity

and emotional cut-off. J. Couns. Psychol. 52(1), 14-24.

Wei, M., Russell, D. W., Zakalik, R. A., 2005b. Adult attachment, social self-efficacy, self-

disclosure, loneliness, and subsequent depression for freshman college students: a

longitudinal study. J. Couns. Psychol. 52(4), 602-614.

40
Wei, M., Heppner, P. P., Russell, D. W., Young, S. K., 2006. Maladaptive perfectionism and

ineffective coping as mediators between attachment and subsequent depression: A

prospective analyses. J. Couns. Psychol. 53, 67-79.

Weinberger, D. A., Schwartz, G. E., Davidson, R. J., 1979. Low anxious, high anxious, and

repressive coping styles: psychometric patterns and behavioral and physiological responses

to stress. J. Abnorm. Psychol. 88, 369-380.

Wells, A., 2009. Metacognitive therapy for anxiety and depression. Guilford Press, NY.

Wells, A., Matthews, G., 1994. Attention and emotion: A clinical perspective. Erlbaum, UK.

Wells, A., Matthews, G., 1996. Modelling cognition in emotional disorder: The S-REF model.

Behav. Res. Ther. 34, 881-888.

Wenzlaff, R. M., Wegner, D. M., 2000. Thought suppression. Annu. Rev. Psychol. 51, 5991.

Zaremba, L.A., Keiley, M.K., 2011. The mediational effect of affect regulation on the relationship

between attachment and internalising/externalising behaviours in adolescent males who

have sexually offended. Child. Youth Serv. Rev. 33, 1599-1607.

Zuroff, D. C., Fitzpatrick, D. K., 1995. Depressive personality styles: Implications for adult

attachment. Pers. Indiv. Differ. 18, 253-365.

41
 Anxious Hyperactivating
Attachment strategies

Depressive
symptoms

Avoidant Deactivating
Attachment strategies

Figure A.1: Our hypothesised model of the theoretical links between attachment, emotion
regulation and depressive symptoms

42
 Table A.1: Summary of Papers Included in Systematic Review
No. Authors Year Title: Country Sample Design Attachment Mediator Outcome Findings Qualit
Measure y
measures measures rating

Studies Utilising Adolescent Samples (in alphabetical order)

1 Brenning, 2012 Attachment and Belgiu Study 1: 125 Cross Experiences in Emotion Childrens Study 1: 14
Soenens, depressive m males and sectional Close Regulation depression Emotion
Braet & symptoms in 214 girls, 12- Relationships- Inventory Inventory regulation (Rated
Bosmans middle childhood 14 years of Analysis: Revised (Roth, (CDI; was did not Poor)
and early age (M= 12.6 Structural (ECRR; Assor, Kovacs, mediate
adolescence: years, SD = Equation Fraley, Waller Niemiec, 1985)
Testing the validity 0.67) Modelling & Brennan, Ryan, & Study 2:
of the emotion 2000) Deci, Suppression
regulation model of Study 2: 292 2009). partially
attachment males and mediated the
438 females, Mediators: relationship
aged 8-14 between
Suppression
years (M= 12 and avoidant
years, SD = dysregulation attachment
1.23) and
depression.
Dysregulation
was a partial
mediator for
anxious
attachment
and
depression

43
2 Cooper, 1998 Attachment styles, US 2,011 male Cross Attachment Various The Brief Anxious- 17
Shaver & emotion regulation and female sectional Questionnaire measures Symptom ambivalent,
Collins and adjustment in adolescents (Hazan & of risky or Index (BSI; poor emotion (Rated
adolescence approximatel Analysis: Shaver, 1990) problem Derogatis regulation Fair)
y half male MANCOVA behaviours & and highest
and half Melisaratos symptom
female exact Mediators: 1983) levels
figures not Risky
reported. 13- behaviours
19 years, (M=
16.7 years,
SD= not
reported)

3 Kobak & 1995 Emotion regulation US 42 mother- Cross- Adult Dysfunctional Dimensions Adolescent 14
anger scale & of
Ferenz- and depressive adolescent sectional Attachment communicative attachment
Gillies symptoms during pairs. Interview assertiveness
Depression was not (Rated
adolescence: A Regression (George, Profile for measured. Poor)
Adolescents model Adolescents
functionalist Kaplan, & (Harter,
perspective 13-19 years Main, 1986)& Marold &
(M= 14.7 Attachment Q Nowakowski
years, SD = sort (Kobak, , 1987)
0.98) Cole, Ferenz-
Gillies, &
Fleming,
1993)

4 Kullik & 2013 Attachment to Germany 248 Cross- A German The German For female 14
Petermann Parents and Peers adolescents sectional short version Regulation version of participants,
as a Risk Factor for of the of Emotion the Center results (Rated
Adolescent 12-17 years Hierarchical Inventory of Questionn for reveal a full Poor)
Depressive multiple
Parent and aire (REQ; Epidemiol mediational
(M = 14.41 regression
Disorders: The Peer Phillips& ogical effect for
years, SD = analyses and
Mediating Role of bootsrapping Attachment Power Studies- internal-
Emotion 1.39) (IPPA; 2007) Depressio dysfunctiona
Regulation Armsden & n scale l
Greenberg) (CES-D; emotion
German regulation.
version, For male
44
 Allgemein participants,
e the relation
Depressio between
nsskala, attachment
ADS; to parents
(Hautzinge and
r, Bailer, depressive
Hofmeiste symptoms
r, Keller; was partially
2011) mediated by
internal- and
external-
dysfunctiona
l emotion
regulation.
5 Torquati & 1999 Attachment as an US 73 female Cross Collins and Child and Beck Insecure 12
Vazsonyi organisational adolescent sectional read (1990) adolescent Depression attachment
construct for affect, students, 18- Analysis: problem Inventory associated (Rated
appraisals, and 22 years solving short form with poor Poor)
coping of late Multiple inventory (Beck et al emotion
adolescent females. (M= 20.6 Hierarchical Gamble 1961) regulation
years, Linear (1993) and
SD=not Regression depressive
reported) Emotion symptoms
regulation
measured

6 Zaremba & 2011 The mediational US 62 male Cross- Attachment Emotion Youth Self- Emotion 14
Keiley effect of affect incarcerated sectional Questionnaire Regulation Report regulation did
regulation on the adolescents (Collins and Checklist (YSR; not mediate (Rated
relationship 13-19 years Analysis: Read 1990) (ERC; Achenbach the Poor)
between Shields & 1991) relationship
attachment and (M= 15.8 Path Cicchetti, between
internalizing/extern years, SD= analysis 1997) insecure
alizing behaviours 1.72) attachment
in adolescent males Mediator: and
who have sexually emotion internalising
offended regulation symptoms

45
7 Ruijten, 2011 The Mediating The 455 Cross The short The The Beck Full 13
Roelofs & Role of Netherla adolescents sectional version of the Ruminative Depressio mediation
Rumination in the nds Inventory of Response (Rated
Rood age range n was found
Relation Between 1218 years Analysis: Parent and Scale Inventory- for Poor)
Quality of Peer (RRS; II (BDI-II- rumination
Attachment (M = 14.3 Bootstrappin Attachment Nolen- NL; Beck in the
Relations and years, SD = g (IPPA; Hoeksema et al. 1996; relation
Depressive 1.3) Armsden and and Dutch between
Symptoms Greenberg Morrow version: communicati
in Non-Clinical 1987: Dutch 1991; Van der on with
Adolescents version: Muris Dutch Does, peers and
et al. 2001) version: 2002) depressive
Raes et al. symptoms,w
2003) hereas
partial
mediation
for
rumination
was found in
the
relations
between
parental
trust and
alienation
from peers,
and
symptoms of
depression

Studies Utilising Adult Samples (in alphabetical order)

1 Burnette, 2009 Insecure US 221 students, Cross Experiences in Rumination Center for SEM 11
Davis, attachment and 141 women, sectional Close -Reflection Epidemiolo detailing a
Green, depressive 80 men, Relationships- Questionna gic Studies path from (Rated
Worthington symptoms: The Structural Revised ire (RRQ; Depression anxious
Jr, & mediating role (M=19.30 Equation (ECR-R; Trapnell & Scale attachment, to Poor)
Bradfield of rumination, years; SD = Modelling Fraley, Campbell, (CES-D) depression,
empathy, and Waller, & (Radloff, mediated by
46
 forgiveness 15.85) Brennan, 1999) 1977) rumination
2000)

2 Garrison, 2012 Disentangling the US 121 students, Cross Experiences in Distress Inventory Attachment 18
Kahn, Sauer effects of 96 women, sectional Close Disclosure of avoidance
& Florczak depression 25 men Relationships Index Depression negatively (Rated
symptoms and Analysis: Scale (ECRS; (DDI, and associated Fair)
adult attachment on (M = 19.76 Multi-level Brennan et al., Kahn & Anxiety with
emotional years, SD= modelling 1998) Hessling, symptoms disclosure
disclosure 1.72) 2001). (IDAS; and
Multiple Watson et disclosure
Hierarchical Mediator: al 2007) negatively
Regression associated
disclosure with
depressive
symptoms.
Attachment
anxiety
associated
with intra-
individual
disclosure

3 Kobak & 1988 Attachment in Late US 53 students, Cross- Adult Q-sort Hopkins Anxiously 12
Sceery adolescence: 27 female and sectional Attachment Measure Symptom attached
Working Models, 26 male, Interview Checklist- groups with (Rated
Affect Regulation, MANOVA (George, Ego- 90 (HSCL- low ego- Poor)
and (M=18.2 Kaplan & resiliency 90) resiliency
Representations of years, SD= Main 1985) measured (Derogatis, reported more
Self and Others not reported) 1977) distress than
those in
dismissing
and secure
attachment
groups

4 Lopez, 2001 Adult Attachment US 55 students. Cross Experiences in The Depression Attachment 11
Mauricio, Orientations and (17 males, 38 sectional Close Problem- Adjective styles
Gormley, College Student females) Relationships Focussed Checklist, associated (Rated
Analysis:
47
 Simko, & Distress: The (M= 21.75 Regression Scale (ECRS; Style of Forms F with coping Poor)
Berger Mediating Role of years, SD= Brennan et al., Coping and G and
Problem Coping 4.74) 1998) (PF-SOC (DACL; psychological
Styles Heppner et Lubin, distress
al., 1995). 1965)

Mediator:

Coping
measured

5 Lopez , 2002 Adult attachment US 127 students, Cross ECR (Brennan Splitting Hopkins Self- splitting 15
Mitchell & orientations and 36 male, 91 sectional Clark and Index (SI; Symptoms and self-
Gormley college student female Analysis: Shaver 1998) Gould, Checklist concealment (Rated
distress, test of a Regression Prentice, & (HSCL; mediated the Fair)
mediational model (M= 19.96 Ainslie). Derogatis, relationship
years, SD= The Lipman, between
1.53) Problem- Ricklets, anxious
Focussed Uhlen- attachment
Style of huth, & and distress.
Coping ( covi, 1974) Avoidant
PF-SOC attachment
Heppner et did not have
al., 1995) an association
with distress
Mediators:

Self-splitting

Self
concealment

48
6 Marganska, 2013 Adult Attachment, US 284 students, Cross RSQ Difficulties Beck Perceived Rated
Gallagher & Emotion 230 female sectional (Griffin & in Emotion Depression inability to (12
Miranda Dysregulation, and and 54 male, Bartholomew, Regulation Inventory, generate Poor)
symptoms of Structural 1994) Scale Second effective
Depression and (M=20.5, SD Equation (DERS; Edition emotion
Generalised = 4.8) Modelling Gratz & (BDI-II; regulation
Anxiety Disorder Roemer, Beck, strategies
2004) Steer, & mediated the
Brown, relationship
1996) between
insecure
attachment
and both
depression
and GAD
symptoms

7 McCarthy, 2006 Preventive US 390 students, Cross Parental Negative Hopkins Association 15
Lambert, & Resources and 267 female, Sectional Attachment Mood Symptom of insecure
Moller Emotion 123 male Questionnaire Regulation Checklist attachment (Rated
Regulation Analysis: (PAQ; Kenny, Scale (HSC; with emotion Fair)
Expectancies as (M: 21.04 1987, 1990) (Catanzaro Greene, regulation,
Mediators Between years, Structural Inventory of & Mearns, Walkey, stress
Attachment and SD=2.34) Equation Parental and 1990). McCormic, symptoms
college Students Modelling Peer & Taylor, and stress
Stress Outcomes Attachment Mediator: 1988) produced
(IPPA; emotions
Armsden & Emotion
Greenberg, regulation
1987, 1989) expectancy

8 Mikulincer 1995 Attachment styles Israel 120 students, Experimental Based on Taylor Idiosyncratic Anxious- 12
and repressive 78 women Hazan & Manifest measure of ambivalent
49
 & Orbach defensiveness: The and 42 men, Design Shaver;s Anxiety mood attached (Rated
Accessibility and 19-27 years measure Scale people Poor)
Architecture of (M= not Analysis: (1987), with Marlowe- reported
affective memories reported, ANOVAs additional Crowne being unable
SD= not rating scale by Social or unwilling
reported) Mikulincer et Desirability to repress
al (1990) Scale negative
(Weinberge affect and
r, Schwartz thoughts
&
Davidson,
1979)

Repression
measured

9 Wei, 2003 Perceived Coping US 515 students, Cross Adult Problem- Beck Perceived 15
Heppner & as a Mediator 349 women, sectional Attachment Focused Depression coping was a
Mallinckrod Between 165 men. 18- Analysis: Scale (AAS, Style of Inventory mediator to (Rated
t Attachment and 41 years, Structural Collins and Coping (BDI; attachment Fair)
Psychological (M=18.93 Equation Read 1990) measure Beck, and
Distress: A years, SD= Modelling Ward, psychological
Structural Equation 2.26) (PF-SOC; Mendelson, distress
Modeling Heppner, Mock &
Approach Cook, Erbaugh,
Wright, & 1961)
Johnson,
1995)

Mediator:

Perceived
coping

50
10 Wei, 2006 Maladaptive US 372 students, Longitudinal Experiences in Problem- Center for Ineffective 19
Heppner, Perfectionism and 219 female, Analysis: Close Focused Epidemiolo coping
Russel & Ineffective Coping 153 men. Structural Relationships Style of gic mediates the (Rated
Young as Mediators Equation Scale (ECRS; Coping Studies relationship Fair)
Between (M=20.01 Modelling Brennan et al., measure Depression between
years, SD= 1998) Scale insecure
Attachment and 1.07) (PF-SOC; (CES-D; attachment
Future Depression: Heppner, Radloff, and
A Prospective Cook, 1977) depression
Analysis Wright, &
Johnson,
1995)

Mediator:

Perceived
coping

11 Wei, Russell 2005 Adult attachment, US 308 students, Longitudinal Experiences in The Center for Avoidant 20
& Zakalik social self-efficacy, 125 male and design Close Distress Epidemiolo attachment
self-disclosure, 183 female, Analysis: Relationships Disclosure gical and (Rated
loneliness and 18-20 years. Structural Scale (ECRS; Index (DDI Studies depression is Good
subsequent Equation Brennan et al., Kahn and Depression mediated by )
depression for (M=18.31 Modelling 1998) hassling Scale comfort with
freshman college years, SD= 2001). (CES-D; self-
students: a 0.47) Kohout, disclosure
longitudinal study Comfort Berkman,
with self- Evans, &
disclosure Cornoni-
measured Huntley,
1993)
UCLA
Loneliness
Scale
(Russell,
1996)

12 Wei, Vogel, 2005 Adult Attachment, US 229 students, Cross Experiences in Differentiat Depression Emotional 15
Ku, & Affect Regulation, 148 women, sectional Close ion of Self Anxiety cut-off,
Negative Mood, 70 men, 11 Analysis: Relationships Inventory and Stress mediated the (Rated
51
 Zakalik and Interpersonal participants Structural Scale (ECRS; (DSI; Scales relationship Fair)
Problems: The did not Equation Brennan et al., Skowron & between
Mediating Roles indicate Modelling 1998) Friedlander (DASS;Lo avoidant
gender. , 1998) vibond & attachment,
of Emotional Lovibond, negative
Reactivity and 18-43 years Mediators: 1995) mood and
Emotional Cutoff (M= 19.88 interpersonal
years, SD= Emotional problems.
2.26) cut-off
Emotional
Emotional reactivity,
reactivity mediated the
relationship
between
anxious
attachment,
negative
mood and
interpersonal
problems

Key:

M = mean age

SD = Standard Deviation

MANCOVAS = Multivariate analysis of covariance

52
Figure A.2 Paper Selection Flow Chart
Papers identified through search in EMBASE and 100 papers
PsycInfo: n= 2087 excluded

Papers after exclusion of papers that were not written 1866 papers
in English excluded
n=1987

Papers after the exclusion of papers that were not 104 papers
relevant to the review according to paper titles and excluded (i.e. for
abstracts not assessing all
constructs)
n= 121

5 papers identified from full-text review of all papers

assessed for eligibility

n=17+5

Eligible papers that met inclusion criteria 3 papers

excluded using
n=22 child samples

Eligible papers

n=19

53
Acknowledgements

None

Conflict of interest

The authors declare no conflict of interest.

Contributors

All authors decided on the focus of the review and the methodology. SM conducted the systematic
literature search and initial analysis. The first draft was prepared by SM, checked by all authors and
finalised by AW. All authors contributed to and have approved the final manuscript.

54
Highlights

Emotion regulation was examined as mediator of attachment and symptomatology

Adult studies found emotion regulation was a mediator
Adolescent studies with limitations found inconsistent results
Associations of hyperactivating and deactivating strategies with depression exist
Methodological limitations, multicollinearity and singularity were discussed
Future directions for research were suggested using clinical samples

Role of funding

The review did not receive any funding.

55