Assessing attachment in the school years: The application of the Dynamic Maturational Model of Attachment to the coding of a child attachment interview with community and looked-after children

Abstract

Background:

This article reports on the analysis of a Child Attachment Interview using the Dynamic Maturational Model (DMM) of attachment coding system developed by Crittenden for use with the Adult Attachment Interview (AAI). The aim of the study was to see if the two coders could classify the child interviews using the DMM-AAI approach and produce the range of DMM attachment strategies to be expected from previous research and the literature.

Methods:

Two coders independently classified interviews with 41 children aged between 6 and 13 years with an average age of 9.8 years. In total, 24 of the children were from a local authority middle school (the community children) and 17 were in foster care (looked-after children).

Results:

The full array of DMM strategies was identified, with significant differences between the community and looked-after children in terms of attachment security and lack of resolution of loss and trauma. There was 100% agreement between coders on secure versus insecure attachment patterns, a Kappa of .910 for the full range of DMM attachment strategies and Kappas of between .655 and .773 for unresolved loss, trauma and depression. Discussion focuses on the strengths and deficits of the use of the DMM compared with other published work on child attachment interviews, the use of interviews to assess post-traumatic stress disorder in children and the implications of defensive attachment strategies for services offered to looked-after children.

Keywords

Attachment interview middle childhood validity dynamic maturational model foster children PTSD

Introduction

The best-validated attachment assessments lie at opposite poles of development, starting with observation of infants (Ainsworth, Blehar, Waters, & Wall, 1978) and preschool children (Cassidy & Marvin, 1992; Crittenden, 1992; 2004) and then moving rapidly up to interviews with adults (Main & Goldwyn, 1984–1994; Crittenden, 1998–2008; Crittenden & Landini, 2011). Assessing attachment in the school years has proved something of a lacuna, and currently no one procedure has been established as the optimum approach. The reasons for this are various, including the lack of comparable measures against which to validate a new system of analysis and an imperfect understanding of what actually constitutes the main indices of attachment behaviour in this age group (Mayseless, 2005).

To add to the difficulties there are two overlapping yet distinct models of attachment theory available, which can be termed the mainstream ABCD model (Cassidy and Shaver, 2008) and the Dynamic Maturational Model (DMM) of attachment and adaptation (Crittenden, 2006; 2008), which is the one employed here. Both models were developed from Mary Ainsworth’s original work with infants aged 11–15 months which provided both a procedure for assessing attachment (the Strange Situation) and the ABC notation to describe three types of infant attachment behaviour: Type A, usually termed insecure avoidant; Type B, balanced, comfortable or secure; and Type C, for insecure ambivalent or passive-resistant (Ainsworth et al., 1978). The ABCD approach has made substantial progress in identifying both disorganised (Type D) behaviour in maltreated infants (Main & Solomon, 1986, 1990) and the mental representation of attachment behaviour in older children (Main, Kaplan & Cassidy, 1985). The latter opened the way for the construction of the Adult Attachment Interview (AAI), which uses a complex system of discourse analysis to identify the different ways in which adults process information about the care they received as children (Main & Goldwyn, 1984–1994). With regard to notation, the Main and Goldwyn replaces the infant ABC categories with the following adult equivalents: Dismissing (Ds), Secure/free/autonomous (F) and preoccupied/entangled (E), together with Unresolved and Cannot be Classified that map the infant D category.

The DMM differs from the ABCD model in a number of crucial respects. In particular, while the ABCD approach continues Bowlby and Ainsworth’s emphasis (Bowlby, 1980; Ainsworth et al., 1978) on the importance of security in attachment relationships, the DMM focuses on adaptation to danger, and so most of what the ABCD model terms disorganised behaviour is reframed by the DMM approach as an organised response to severe threat (Crittenden, 2000, 2006; Farnfield, Hautamäki, Nørbech, & Sahhar, 2010).

A number of studies have used the Main and Goldwyn AAI coding system to code interviews with school-aged children (e.g. Ammaniti, Van IJzendoorn, Speranza, & Tambelli, 2000; Borelli et al. 2010a; Borelli, David, Crowley, & Mayes, 2010b; Shmueli-Goetz, Target, Fonagy, & Datta, 2008; Steele & Steele, 2005). This is of considerable importance in itself, as they demonstrate that by the age of 7 years the human mind has already organised the basic linguistic constructs with regard to attachment which are found in adulthood. Commensurate with the AAI, the coherence of children’s discourse (their speech patterns) is assumed to be a window on how they think (process information) about relationships and threatening situations. For example: psychologically defended or dismissing children (in DMM terms Type A attachment) try to minimise the impact of rejection or the unavailability of attachment figures; those in Type C are preoccupied with problems in relationships, and those in Type B show increasing ability to reflect on both their own and other’s contributions to interpersonal difficulties and maintain coherence even when talking about distressing experiences.

The empirical base for the DMM-AAI is currently modest, and there are no DMM-based studies of interviews with school-age children. With these limitations in mind it is also apparent that the ABCD model has encountered a number of problems; chief of these is the question about what happens to ‘disorganisation’ post infancy. So far, disorganised attachment which persists into adulthood is either treated as unclassifiable in the AAI or identified in terms of the lack of resolution of loss and/or trauma (Hesse, 2008). Previous studies using a Child Attachment Interview (CAI) have employed either a disorganised category (Shmueli-Goetz et al., 2008; Target, Fonagy, & Shmueli-Goetz, 2003) or the AAI’s unresolved loss or trauma / cannot classify (Ammaniti et al., 2000) but, with the exception of Borelli and colleagues who have conducted a specific study of disorganised attachment in this age group (Borelli et al., 2010b), they produced too few examples to generate useful information. However, work on narrative story stems has identified organised strategies in 6-year-old children assessed as disorganised in infancy (Solomon, George & De Jong, 1995). None of these studies produced data about lack of resolution of trauma (post-traumatic stress disorder (PTSD)/ Developmental Trauma Disorder (DTD)).

In addition, when applied to CAIs (Ammaniti et al., 2000; Shmueli-Goetz et al., 2008) and narrative stems (Green, Stanley, Smith, & Goldwyn, 2000) the ABCD model has identified a disproportionately low number of children using the Type C anxious-ambivalent/preoccupied strategy, and sometimes assesses a disproportionately high numbers of children from at-risk backgrounds as securely attached.

The dynamic maturational model of attachment

The DMM attachment strategies for the school years are given in Figure 1.

Figure 1.

A dynamic maturational model of patterns of attachment in school age (with permission from Patricia M Crittenden).

Crittenden organises her model around two behavioural dimensions: cognition and affect. By cognition she simply means learned behaviour based on previous experience of cause and effect. This appears in children’s speech as the rules about how they should behave; for example, ‘My Mum gets worried when I am late’ or ‘If you get in Dad’s way he shouts’. The Type A and A+¹ strategies involve excessive use of cognition in order to minimise the effects of stress-related feelings. In Type A1-2 threats to children’s safety are usually low, and children deploying this strategy tend to live in environments which are protective but discourage protracted displays of emotion, in particular anger and the desire for comfort.

The higher Type A+ sub-patterns are referred to as compulsive because children not only inhibit behaviour forbidden by adults, but shape their behaviour to meet the demands of the outside world. Type A3 refers to compulsive caregiving or role reversal, which is common among neglected children or where parents are depressed or otherwise pull back when their children approach for comfort (Bowlby, 1980; Crittenden, 1992; George & Solomon, 1999). The desire for nurture thus becomes a forbidden emotion for the child, who responds by becoming the psychological parent to their attachment figure(s) in order to elicit whatever protection is available. Type A4 refers to the robotic compliance with adult demands met with in physically abused children (Crittenden & DiLalla, 1988).

Children using Type C strategies face the opposite set of problems to those in Type A. Their attachment figures are unpredictable and tend to pay attention to negative (from the parent’s perspective) child behaviour, thus reinforcing behaviour which is overtly forbidden (but may be covertly encouraged), and the family structure is typically enmeshed, with children sometimes becoming the unwitting pawns in adult relationship struggles. Hence cognition (what causes what) is not a good guide to behaviour, and children develop forms of affective logic (typically displays of anger or extreme vulnerability and helplessness) to coerce parents and, by the school years, other adults, into maximising attention on the child. The function of the child’s behaviour is to render attachment figures predictable (Crittenden, 2008).

Like the Type A1-2 strategy, Type C1-2 does not pose a threat to long-term psychosocial development and is likely normative for many cultures. When the child is anxious or there are unresolved problems in the family the child emphasises anger (Type C1) or the desire for comfort and nurture (Type C2), sometimes alternating the two strategies to maximise parental attention. The Type C3 (aggressive: attachment hot bloods) and C4 (feigned helpless) patterns exaggerate one affective state at the expense of the other (anger hides vulnerability or vice versa) and generalise the strategy out of home, with the result that the angry brigade tend to be marked out as having conduct and/or attention deficit disorders. By the end of the school years some of these children have reorganised into a much cooler use of affect which includes the ability to deceive others as to their intentions (false cognition) and where behaviour is focused on punitive revenge (Type C5) or the use of seduction, sometimes sexual but more often extreme risk-taking behaviour, to present the self as helpless and in need of rescue (Type C6). All people in Type C actually perceive themselves as victims, whereas those in Type A over-attribute responsibility to themselves for the bad things that befall them (Crittenden, 2008).

Some children from complex environments show A/C combinations which may be a response to levels of threat within the same relationship (e.g. C3 when mother is well and A3 when she is depressed) or between relationships (C1 with a benign but inconsistently attentive mother and A4 with her angry boyfriend).

For sake of completeness the full range of strategies available in adulthood is given in Figure 2. On the A side of the model A5 refers to compulsive promiscuous behaviour, which functions to provide social and/or sexual contact without the risk of intimacy, and A6 to compulsive self-reliance (Bowlby, 1980) as a result of repeated failures to protect the self using the lower Type A strategies. Type A7 is associated with severe and ongoing abuse as a child, in which the adult now idealises an attachment figure who was, in reality, life threatening. The externally assembled self (Type A8) is the most extreme of the Type A+ patterns, and involves absence of information about large periods of childhood together with an attempt to assemble a self-identity from professional opinion, diagnoses and official records. The A8 pattern is associated with severe abuse as a young child and also multiple foster placements. Although school-age children are not able to use the A5-8 defences strategically, their development is clearly discernible in interviews with children and were prominent in the data from the looked-after children in this study.

Figure 2.

A dynamic maturational model of patterns of attachment in adulthood (with permission from Patricia M Crittenden).

Types C7-8 refer to free-floating dread where everyone and everything is perceived as a source of threat, and has proved rare in the CAIs conducted so far.

Children in Type B are both protected and comforted and have opportunities, not available to those in the insecure patterns, to explore the self and their relationships with others. By the school years they are predicted to be able to repair problems in relationships and to be on the optimum pathway to what Fonagy and colleagues call mentalising: the capacity to attend to mental states in oneself and others, especially those mental states which differ from one’s own (Allen, Fonagy, & Bateman, 2008).

Post-traumatic stress disorder

There is a growing interest in increasing the accuracy of the diagnosis of PTSD/DTD in children together with the construction of diagnostic categories and the use of rating scales (Hawkins & Radcliffe, 2006; Hoge, Austin, & Pollack, 2007; van der Kolk, 2005). A limitation of the latter is that they can only be used when a potentially traumatic event has been identified, which means that pervasive trauma, of the sort experienced by many children who are taken into local authority care, may remain undetected. The DMM raises the question of the degree to which children are able to organise defensive behaviour around pervasive trauma; for example, child neglect can produce role reversal (Type A3 compulsive caretaking of attachment figures) (Crittenden, 2006, 2008; Marvin, 2003) in which exposure to ongoing threat is actually subsumed into the defensive strategy. When this is the case, a diagnosis of PTSD seems redundant.

PTSD is assessed in the DMM when lack of resolution to trauma actually derails the strategy; in the neglect example, when compulsive caregiving does not function to keep the parent’s attention, the child might then exhibit PTSD symptoms. One basis for the behaviours in children diagnosed with PTSD may actually be reversals in the usual defensive strategy, such as dismissed trauma in a Type C+ strategy or preoccupied loss in a Type A+.

The DMM-AAI also has discourse markers for modifiers which reflect the failure of interpersonal self-protective strategies. Crittenden identifies these as depression (the awareness that the strategy is not working; a common feature of PTSD in adults), disorientation (incompletely described, but referring to confusion as to where information about relationships actually comes from, with a concomitant over-attribution to the self as a source of all information) and intrusions of forbidden negative affect in Type A+ strategies. The latter are typically a sudden rush of comfort-seeking behaviour in a compulsive caretaking child for whom comfort is forbidden, or outbursts of rage in an over-controlled, compulsive compliant (Type A4 strategy) child. In both cases the child is out of their own control and unable to curb potentially anti-social behaviour (sexual in the case of forbidden comfort and violence in the case of forbidden anger) (Crittenden & Landini, 2011).

Unlike a rating scale, the use of an attachment interview, or narrative story stems, offers the clinician a wider net to cast over the child’s autobiography with more chance of identifying traumas that need professional attention.

The DMM and Main and Goldwyn AAI notations

In terms of comparing the two systems, the DMM A 1-2 corresponds with the Main and Goldwyn (M&G) Ds 1-2; DMM B 1-5 with the M&G F1-5; DMM C1-2 with the M&G E 1-2; the DMM C5 with M&G Ds2, and the DMM unresolved loss in an anticipated form with the M&G Ds4. The M&G E3 category corresponds with the DMM C3-8 but also some DMM A3-8 patterns, and the M&G Cannot Classify “often A/C, but … some A3-8 and C3-8,” (see Crittenden and Landini, 2011, p. 388–389).

Methodology

The aim of the study was to see if application of the method of discourse analysis developed for the DMM-AAI (Crittenden 1998–2008; Crittenden and Landini, 2011) could produce the expected array of DMM patterns found in the theoretical literature and previous research.

Participants

Interviews were conducted with 41 children (19 girls and 22 boys) aged 6–13 years with an average age of 9.8 years. Twenty-four children were drawn from a larger data set (n = 32) of interviews with children aged 7–11 years and are termed the community children. A further 17 were drawn from a set of interviews with children in foster care and are called here the looked-after children. The community children were interviewed in 2000 and the looked-after children between 2000 and 2007.

The community sample came from a local authority middle school with pupils aged 7–11 years, located in the centre of a large town in the South of England; 32% of the children were on the school’s register of special educational needs, and attainment on entry was generally below average (2002 Ofsted report for the school). All the parents were approached and were told we wanted to interview children who had lived continually with at least one of their parents, had not had any treatment for emotional or psychological problems and had never been on the local authority child protection register. The 32 consents represented 8% of the school roll and included two children who had had treatment for emotional problems: one for an eating disorder and the other for needle phobia. The sample was biased in favour of children whose parents had either separated (25% (n=8) of the children had parents who were separated, divorced or had never lived together) or worked in a social care occupation (just under 33% (n=11) of the children had at least one parent working for health, education or social services). Three of the community children were on the school register for special educational needs. This was a densely populated town and all the children lived within a mile of the school.

The looked-after children were drawn from an initial set of 42 research interviews that formed part of a PhD thesis (Farnfield, 2003) supplemented by a further set of 129 interviews collected on a convenience basis over a number of years, including data collected by students on the post-qualifying child care award at the University of Reading and interviews conducted as part of clinical assessments. The looked-after children all had complex histories, many of them involving severe abuse and/or neglect while with their birth families and several changes of placement once in the care system.

Financial constraints limited the number of interviews we could code, so a subset of 24 children from the community group was chosen to represent as far as possible both sexes at the ages of 7–11. It also included the two children who had had psychological treatment. The 17 looked-after children were chosen to provide a match with the community children in terms of age and sex.

While considerable historical and professional material was available on the looked-after children, the only additional information about the community group was a brief screening questionnaire and the interviews themselves. In terms of socio-economic background, the community children were more advantaged than the looked-after children. Using the Registrar of Occupations, the highest social class for either parent for the community children was: class 1, 8%; class 2, 25%; class 3, 52%; class 4, 12%; and class 5, 3%; compared with the looked-after children: class 3, 23%; class 4, 36%; and class 5, 41%.

Ethical approval

Ethical approval was given by the school governors for the community sample, the social worker for the at-risk children and the University of Roehampton ethics committee reference PT 09/008.

Measures

All the children gave a CAI (Farnfield, 2001, 2009) and completed a set of narrative story stems designed for this age group (Farnfield, 2001). The results of the narrative story stems will form the basis of a separate paper. The current version of the CAI is given at Appendix 1.

Procedure

The interviews with the community children were conducted in a private room at their school by the author. Each interview took about 40 minutes to complete, with the children coming from class during school time. One girl came with a friend but all the others were interviewed alone. Of the looked-after children, eight were interviewed by the author and the other nine by social workers from the local authority who had care of the child. Six of the looked-after children were interviewed in their foster homes; one girl was interviewed in her children’s home, one boy in a family centre and another boy was in secure accommodation. None of the community children were known to the interviewer, but four of the looked-after children were interviewed by a social worker who knew them quite well.

All the interviews were tape recorded and then transcribed following the verbatim transcription of hesitations, pauses and other dysfluencies established for the AAI. Where possible, references were made to non-verbal behaviour such as smiles, laughs, coughs and so forth.

All the children, one of their parents or, in the case of some looked-after children, a social worker, gave consent to the interviews following written and verbal explanation of the study.

Coding

The interviews were coded by the author and by a second coder who, except for the age and sex of the interviewees, was blind to the identities of the children and to the nature of the two populations. Both coders have research-level reliability in the DMM-AAI.

Coding was conducted using our knowledge of the DMM-AAI without recourse to a separate manual or coding system tailored to this age group. Interviews were classified using the full range of DMM attachment strategies together with the modifiers and lack of resolution of loss and trauma (PTSD).

Results

Inter-rater reliability

There was 100% agreement between the two coders in terms of secure Type B versus the insecure patterns, and 88% agreement on the insecure interviews giving a Kappa for the full range of DMM strategies of .910. The percent agreement for depression was 66% (Kappa .773); unresolved loss 80% (Kappa .723) and unresolved trauma 60% (Kappa .655). Agreement for unresolved loss or trauma was 60% (Kappa .655). The disorientated modifier and intrusions of forbidden negative affect were each identified once by a single coder with no agreement between coders.

No agreement could be reached for two of the children with regard to attachment strategy, both of whom were from the looked-after group, and were entered in the data as A/C.

The community children

Table 1 gives the distribution of attachment strategies for both groups of children; 46% (n=11) of the community interviews were rated secure Type B, 12% (n=3) Type A1-2 and 33% (n=8) Type C1-2. Thus 92% (n=22) of the community interviews were rated in the normative range of security/insecurity, with just two rated in the more extreme C3-4 pattern. One of the interviews rated C4 came from one of the two children who had received psychological treatment – this was for an eating disorder; the interview from the other child (treated for needle phobia) was rated Type B4.

Table 1.

Distribution of attachment strategies.

Attachment	A7-8	A5-6	A3-4	A1-2	B1-5	C1-2	C3-4	C5-6	A/C	Total
Community				3	11	7	3			24
Looked after	3	3	3					3	5	17
Total	1	5	3	3	11	8	2	3	5	41

There were no significant differences with regard to attachment strategy and age or sex. Compared with the percentages given for non-clinical adult populations when using the Main and Goldwyn ABCD-AAI, there were fewer children in Type B (46% compared with 58%) and Type A (12% compared with 23%) and more in Type C 1-2 (33% compared with 19%) (Bakermans-Kraenburg & van IJzendoorn, 2009). Assuming for the time being that the DMM and ABCD systems are assessing the same behaviours under the same ABC categories, the same differences pertained when the results were compared with the CAI-ABCD studies: they had more children in the dismissing and fewer in the preoccupied category than this study. The Ammaniti and Shmueli-Goetz studies (Ammaniti et al., 2000; Shmueli-Goetz et al., 2008) had higher numbers of secure children, and the Borelli study (Borelli et al., 2010a) about the same (the Borelli community sample had 44% secure).

Table 2 gives the distribution of unresolved loss and trauma, and depression, for the two groups.

Table 2.

Unresolved loss and/ortrauma and depression.

	ULossUtrauma	Dp
Community	6 (25%)	0
Looked after	10 (59%)	6 (35%)
Total	16	6

5 of the community children were unresolved re parental separation.

Parental separation

Seven of the 24 community children had parents who were divorced, separated or had never lived together, and of these five interviews were rated as showing lack of resolution with regard to the break-up of their families or loss of their father. Four out of these five children were boys, and the interview of the one girl was also rated unresolved regarding the death of her grandmother.

All three interviews coded Type A1-2 were with children whose parents were divorced, giving a statistically significant association in the community sample between Type A1-2 attachment and divorce (Pearson χ²; 11.413, df 3, p .01). These same three interviews were also rated unresolved regarding the trauma of divorce/separation, and in two cases lack of resolution took a preoccupied form suggesting the child’s usual Type A coping strategy was not functioning to protect them.

Two of the community interviews were rated unresolved with regard to ongoing bullying, and none of the community interviews were rated as showing depression.

The looked-after children

The looked-after children have a very different profile. None of the interviews were rated secure or in the typical Type A1-2 and C1-2 range of insecurity. Nearly 53% (n=9) of the interviews were judged to be in the compulsive A3-8 range, with two of these showing strong signs of the Type A8 externally assembled self; 18% (n=3) were judged to be in Type C5-6 and 29% (n=5) in Type A/C. In three of the A/C cases the strategy seemed to function as a split between foster and birth mother, involving idealisation of one and angry disparagement of the other. This distribution of strategies is very close to that found in a small study by Gogarty (2002), who used the DMM-AAI with 16 adults who had been in long-term foster care as children (Type B 6%, C3-4 19%, A4-5 45% and A/C 30%).

Interviews from nine children were rated as showing unresolved loss and/or trauma, and of these three were rated depressed and a further three interviews were rated as depressed with no evidence for lack of resolution of loss or trauma. Thus six of the looked-after interviews were judged to show depression, and of these five were in a high Type A+ and one was an A/C.

The losses these children had been exposed to frequently involved abandonment by attachment figures, but also deaths. The traumas were a catalogue of neglect, domestic violence and violence to the child, with some truly horrifying accounts of kidnap, adults threatening adults with guns, and houses being set on fire. One interview was rated unresolved with regard to parental separation because this was a prominent feature of the discourse; in other cases, parental separation must have played a part but was not easily discernible in a wider tapestry of neglect and abandonment.

Limitations of this study

This was a small exploratory study and there were no data to provide support for construct validity in terms of parents’ state of mind with regard to attachment, reflective functioning or styles of caregiving. The looked-after group of children were at the extreme end of the range in terms of the degree of danger they had experienced, and it is important to test this approach against a sample of children with clinical problems who are still living at home with their parents. Although the school from which the community children were recruited had nearly a third of its pupils registered for special educational needs this applied to only 12% of the community sample, and their socio-economic profile was higher than that of the looked-after children and likely higher than average for the school they attended. A further limitation is that, given the complexities of DMM-AAI discourse analysis, we do not know how feasible it will be to train coders in the CAI who are not already conversant with the DMM-AAI.

Discussion

When compared with the ABCD studies, the DMM offers solutions to some problems but poses serious questions regarding differences in the coding systems. First, it identified a significant number of preoccupied Type C interviews but, unlike the ABCD studies, the number of dismissing Type A interviews in the community group was low (n=3, or 12%). Second, by using the organised A+ and C+ DMM strategies, it was able to discriminate between the community and looked-after group without resorting to a disorganised or cannot classify category. Third, it did not classify interviews with children from known at-risk backgrounds as Type B secure. Fourth, it was possible to be specific about traumatic experiences the children had suffered.

With regard to the identification of the preoccupied Type C strategy, one of the issues for the ABCD model has been the reliance on emotional openness regarding negative experiences as an indicator of Type B security (Main et al., 1985; Oppenheim, 1997; Shmueli-Goetz et al., 2008). While the ability to discuss negative feelings is an aspect of security in both systems, in the DMM it also functions as a Type C marker when the speaker (adult or child) exaggerates negative affect in order to portray the self as a victim. Thus many of the interviews judged to be Type C using the DMM will be judged as Type B in the ABCD model.

A second over-attribution of Type B may occur in ABCD codings with regard to interviews which the DMM approach would classify as compulsive Type A+. For example, Shmueli-Goetz and colleagues note that of their clinically referred sample just under 10% of interviews were rated disorganised, compared with nearly 50% in other high-risk samples, and about 25% were rated secure Type B. They suggest that this over-representation of Type B in at-risk groups may be due to high levels of coherence (the hallmark of Type B discourse) around negative interactions with attachment figures, and write: “The child’s coherence might reflect an effort to maintain a good relationship with his or her parents despite current negative interactions” (Shmueli-Goetz et al., 2008, p. 252). This neatly describes the DMM’s use of Type A+ ‘exoneration’: the child’s acceptance or acknowledgement that parents have failed to protect her while absolving the parent from any blame. This functions to retain information about negative events while maintaining a model of the attachment figure as protective.

A third area of disagreement is that dismissing Ds2 discourse in the Main and Goldwyn system includes dismissing other people’s perspective which is coded under Type C5 in the DMM (Crittenden & Landini, 2011).

The identification of specific unresolved traumas and losses is promising, and a CAI could help formulate more child-specific diagnoses of PTSD/DTD. Borelli and colleagues found self-reported depressive symptoms to be above the clinical threshold in children whose CAI was rated disorganised (Borelli et al., 2010b), and the inclusion of the depressed modifier in this study revealed high levels of depression in the discourse of the looked-after children.

Two DMM-AAI modifiers, disorientation and intrusions of forbidden negative affect, were identified once with no agreement between coders. While, in the author’s experience, disorientation is relatively uncommon in adult AAIs, the high number of compulsive Type A+ interviews from the looked-after group would, again based on coding adult interviews, yield more examples of intrusions of forbidden negative affect. Why there was only one is not clear. It may reflect errors in coding, a weakness in the interview format to elicit the material, or that it does not manifest itself in discourse until adulthood.

The identification of Type A5-8 strategies in the looked-after group requires further study regarding whether, as DMM theory suggests, they are in a pre A5-8 form or are actually fully functioning self-protective strategies.

Implications for treatment of looked-after children

Looked-after children are significantly more likely to experience serious psycho-social and mental health problems than any other group of children (McCann, James, Wilson, & Dunn, 1996). In this study none of the interviews with the looked-after children were rated secure, and nearly half were rated Type A+. Although the total looked-after sample is small (only 17 children), the high proportion of compulsive Type A+ interviews conforms to the distribution in the larger data set and to case experience. Further research in this area is urgently needed, but the preponderance of Type A+ makes good theoretical sense in that, as well as abuse, these children had experienced at least one and, typically, several changes of carer, which meant survival was dependent on the child adapting to the demands of other people. This is highly likely to produce compulsive attachment behaviour, whereby the child develops a global strategy designed to mould the self to the perceived expectations of other people, leading to a false self and, in extreme cases, an externally assembled self (Type A8).

Unlike Type A+, a Type C or C+ strategy is intensely interpersonal and demands ongoing relationships with the same attachment figures. Significantly, the three looked-after interviews assessed as Type C5-6 all came from children who had ongoing contact with their birth families.

Children using compulsive strategies are likely to be at greater risk of depression than those in Type C or Type A/C. In the DMM, depression refers to interviews where the child is aware of the problems he faces and that he has no solutions to them. This is more likely in Type A+ because the self is held responsible for managing negative events, and so when the strategy fails self-blame and a lack of self-agency are likely to follow. Conversely, the Type C (+) strategies function to protect the self by blaming others and raising arousal (‘emotional openness’) to motivate action.

This raises important implications for the management and treatment of children in foster care. In particular:

Every change of placement will increase the A+ strategy.

Forms of ‘life story work’ should be devised that enable children to generate their own stories, not the one told to them by adults.

All looked-after children should be screened for PTSD and depression using both a CAI and the available self or parent-rating scales (see Hawkins & Radcliffe, 2006).

Compared with the ABCD model, the DMM offers a wider range of strategies consistent with development and caregiving experience, and is more sensitive to the variations in behaviour found in this age group. However, because it is premised on danger as the environmental condition around which attachment behaviour is organised, it is possible that it over-identifies PTSD.

Footnotes

Appendix 1

CAN YOU TELL ME? - CHILD ATTACHMENT INTERVIEW

For use with children aged 7 to 12 (Abbreviated)

There are extra questions for children in care as follows: the underlined questions 12, 13, 15 and 21. Note that question 11 applies to ALL children

Probes are in italics

1. INTRODUCE THE INTERVIEW

Just for the tape, what’s your name?

And how old are you?

When are you … (next birthday)?

Can you tell me what day it is today?

And what time it is?

NOW TEST THE TAPE RECORDER

2. Can you think very hard and tell me the very first thing that you can remember in your life, and then go on to tell me the story of your life, putting in the things that matter up to now

Start with the first thing you can remember and then go on with the most important things that have happened up until now.

What is your first memory?

Can you remember your first day at school or at nursery?

According to the completeness of the autobiography continue with question 3.

3. Can you tell me who lives here/at home with you?

I’ll write them down.

Is that all the people?

Are there any animals?

Tell me about them.

And do you have a bicycle?

Tell me about it

Here is a chart with houses on it (flow chart)

Starting with the first place can we put down all the places you have lived?

For each house

Do you know where it was?

And who lived there (don’t forget the animals)?

Was that a happy house or a sad house?

Can you say what was happy / sad about it?

Can you give me an example?

4. Now I’d like us to do a chart putting in all the people in your life whom you love or who love you (ECOMAP)

Child or adult complete the ecomap

Think back for a moment, can you tell me something about these people (ON THE CHART)

For example, how do you/did you get on with them

Which of these people would you say you are closest to?

If necessary, probe for: mother, father, siblings, other relatives, brothers and sisters, friends, teachers, others.

How often do you see them?

Is there anyone you would like to see more often?

5. SHOW THE CHILD THE PICTURE

This is a picture of a child in hospital. See?

Now, just suppose this is you in bed there, and you still feel very ill, who do you most want to come and see you?

Anyone else?

6. Has anyone close to you died?

How did you feel about it?

How do you feel about it now?

Did you go to the funeral?

Tell me about that

7. Can you tell me a bit about your Mum and Dad?

What are they like?

Do they worry about you?

Do you worry about them?

Can you think of one really good thing about being with your Mum?

Can you tell me about a time when that happened?

And what about a difficult time with your Mum?

Can you give tell me a time when that happened?

Now let’s do the same with your Dad.

Can you think of one really good thing about being with your Dad?

Can you tell me about a time when that happened?

And what about a difficult time with your Dad?

Can you give tell me a time when that happened?

Tell me about a time you were really angry with your Mum

What do you think your Mum thought about it?

Tell me about a time when you were really angry with your Dad

What do you think your Dad thought about it?

Now tell me about a time when they were really angry with you

Which parent do you feel closer to?

Continue for any other major attachment figures

8. Now I’d like to hear about school².

Tell me about the school you go to now

Do you like going to school?

Tell me about your teacher.

What would your teacher say about you?

Do you think bullying is a problem for children at school?

Has bullying ever been a problem for you?

(If so) What happened?

Did you tell anyone?

Did anyone sort it out?

Have you ever bullied anyone?

9. This bit is about friends. Would you say you have:

A lot of friends

A few friends

Or not many friends at the moment?

Tell me the names of your friends

Do you see them after school/in the holidays?

Do they call for you?

Do you get asked to parties or sleep overs?

Have you got a best friend?

Tell me about him/her?

Can you tell me about a time you fell out or got angry with each other?

What did you feel?

And what do you think your friend was thinking?

10. Can you tell me about a time when you first slept away from home?

Do you remember how you felt about it at the time?

What did you feel like doing?

11. Can we talk a bit about foster care?

What reasons do you think children go into care?

No prompts; leave this question as it is.

12. Tell me about how you first came into care

Who brought you into care?

And why do you think you came into care?

Do you remember how you felt about it at the time?

What did you feel like doing when you came into care?

13. What happens about seeing your (birth) family and relatives?

Would you like to see them more often?

14. Do your Mum and Dad live apart?

(If no) Do you ever worry they might live apart?

(If yes) Do you know why they live apart?

What did you feel about it at the time?

And what do you feel about it now?

Has your Mum or your Dad got a new partner?

How do you get on with them?

15. Let’s talk about the people you live with now

(FOSTER HOME)

What do you call them/want to call them?

Will you stay with them do you think?

Do you want that to happen?

Does your foster carer treat you as a daughter/son do you think?

Is that what you want?

(CHILDREN’S HOME)

What do you call them/want to call them?

Will you stay here do you think?

Do you want that to happen?

Does anyone here treat you like a daughter/son?

Is that what you want?

16. What do you do when you feel upset?

When was the last time you were upset?

What did you do then?

Supposing you felt really very worried about something, who would you tell?

And if you did something wrong, who would be most upset by it?

If people found out who would be most upset?

Have you done anything that really frightened you?

Tell me about that

Tell me about a really happy time

And about a time that was really sad.

17. Do you like being X (child’s age)?

If no - what age would you like to be?

If yes - what age would you like to be if you weren’t X?

Can you say why?

Can you tell me about a time when you were ill?

What happened?

And a time when you were really scared?

What happened?

Does anyone give you cuddles?

Do you like to be cuddled?

Do you like yourself?

Do you think you’re a nice person?

18. How well do you sleep at night?

Do you take a long time to get off to sleep?

Do you wake up in the night?

Do you have dreams?

Are they nice or not nice dreams?

Can you tell me about one of these dreams?

(19. STORY COMPLETIONS HERE)

20. What will happen in your life next do you think?

What do you want to happen next?

Do you think it will?

Optional for clinical use

If we look back over all the things we have been talking about, what needs to change do you think?

And what needs to stay the same?

Would you say you are happier now than you were 6 months ago/ since X happened or less happy or about the same?

Can you say why?

21. Are you pleased you came into care?

22. How do you feel now?

Is there anything else you want to say or ask me about?

For research interviews ask the child to give a pseudonym

Thank you very much.

Acknowledgements

The author expresses his thanks to Gordon Somerville who coded the interviews.

Funding

The research was part funded by a grant from the University of Roehampton, Small Grants Fund, 2009.

Notes

Author biography

Steve Farnfield is a Senior Lecturer and convenor of the MSc in Attachment Studies at the University of Roehampton which he established in 2011. He is a social worker and play therapist and a licensed trainer for the Dynamic Maturational Model of attachment Infant CARE-Index, Preschool Assessment of Attachment and Adult Attachment Interview developed by Dr Patricia Crittenden.

References

Ainsworth

M. D. S.

Blehar

M. C.

Waters

Wall

(1978). Patterns of attachment: A psychological study of the strange situation. Hillsdale, NJ: Lawrence Erlbaum.

Allen

J. G.

Fonagy

Bateman

J. G.

(2008). Mentalizing in Clinical Practice. Washington, DC: American Psychiatric Publishing Inc.

Ammaniti

Van IJzendoorn

M. H.

Speranza

A. M.

Tambelli

(2000). Internal working models of attachment during late childhood and early adolescence: An exploration of stability and change. Attachment & Human Development, 3, 328–346.

Bakermans-Kranenburg

M. J.

van IJzendoorn

M. H.

(2009) The first 10,000 Adult Attachment Interviews: Distributions of adult attachment representations in clinical and non-clinical groups. Attachment & Human Development, 11, 223–263.

Borelli

J. L.

Crowley

M. J.

David

D. H.

Sbarra

D. A.

Anderson

G. M.

Mayes

L.C.

(2010a). Attachment and emotion in school-aged children. Emotion, 10(4), 475–485.

Borelli

J. L.

David

D. H.

Crowley

M. J.

Mayes

L. C.

(2010b). Links between disorganized attachment classification and clinical symptoms in school-aged children. Journal of Child and Family Studies, 19, 243–256.

Bowlby

(1980). Attachment Separation and Loss Volume 2: Separation, Anxiety and Anger. London, UK: The Hogarth Press.

Cassidy

Marvin

R. S.

, with the MacArthur Attachment Working Group. (1992). Attachment organization in pre-school children: Coding guidelines (4th edition). Unpublished manuscript, University of Virginia.

Cassidy

Shaver

P.R.

(Eds.) (2008). Handbook of Attachment, Theory Research and Clinical Applications. New York, London: The Guilford Press.

10.

Crittenden

P. M.

(1992). Quality of attachment in the preschool years. Development and Psychopathology, 4, 209–241.

11.

Crittenden

P. M.

(1998–2008). Patterns of attachment in adulthood: A dynamic-maturation approach to analyzing the Adult Attachment Interview. Unpublished manuscript, Miami, FL, USA.

12.

Crittenden

P. M.

(2000). A dynamic-maturational approach to continuity and change in pattern of attachment. In Crittenden

P. M.

Claussen

A. H.

(Eds.), The organization of attachment relationships: Maturation, culture, and context (pp. 343–357). New York: Cambridge University Press.

13.

Crittenden

P. M.

(2004). The Preschool Assessment of Attachment. Coding manual. Unpublished manuscript, Miami, FL, USA.

14.

Crittenden

P. M.

(2006). A Dynamic-Maturational Model of attachment. Australia and New Zealand Journal of Family Therapy, 27, 105–115.

15.

Crittenden

P. M.

(2008). Raising Parents. Attachment, parenting and child safety. Cullompton: Willan.

16.

Crittenden

P. M.

DiLalla

D. L.

(1988). Compulsive compliance: The development of an inhibitory coping strategy in infancy. Journal of Abnormal Child Psychology, 16, 585–599.

17.

Crittenden

P. M.

Landini

(2011). Assessing Adult Attachment: A Dynamic-Maturational Approach to Discourse Analysis. New York: Norton.

18.

Farnfield

Hautamäki

Nørbech

Sahhar

(2010). DMM assessments of attachment and adaptation: Procedures, validity, and utility. Clinical Child Psychology and Psychiatry, 15(3), 313–328.

19.

Farnfield

(2003). Being in Care: A study of older children’s experiences of local authority care using an attachment and object relations perspective. Submitted in partial fulfilment of the requirements of the University of East London for the degree of Doctor of Philosophy.

20.

Farnfield

(2001, 2009). Can you tell me? Child Attachment Interview for children aged 7–12 years. Versions 1 & 2. Unpublished:The University of Reading

21.

Farnfield

(2001). Narrative stems for use with school aged children aged 7–12 years.

22.

George

Solomon

(1999). Attachment and caregiving: The caregiving behavioural system. In Cassidy

Shaver

(Eds.), Handbook Of Attachment: Theory, Research And Clinical Implications. (pp. 649–670). New York: The Guilford Press.

23.

Gogarty

(2002). Attachment relationships in the triad of foster-care: A retrospective analysis. Thesis submitted in partial fulfilment of the requirement for Doctor of Philosophy, University of Ulster, Coleraine, Ireland.

24.

Green

Stanley

Smith

Goldwyn

(2000). A new method of evaluating attachment representations in young school-age children: The Manchester Child Attachment Story Task. Attachment and Human Development, 2, 48–70.

25.

Hawkins

S. S.

Radcliffe

(2006). Current measures of PTSD for children and adolescents. Journal of Pediatric Psychology, 31(4), 420–430.

26.

Hesse

(2008). The Adult Attachment Interview: Protocol, method of analysis, and empirical studies. In Cassidy

Shaver

P.R.

(Eds.), Handbook of attachment: Theory, research and clinical applications (pp. 552–598). London: Guilford Press.

27.

Hoge

E. A.

Austin

E. D.

Pollack

M. H.

(2007). Resilience: Research evidence and conceptual considerations for post traumatic stress disorder. Depression and Anxiety, 24, 139–152.

28.

Main

Goldwyn

(1984–1994). Adult attachment scoring and classification system. Unpublished manuscript, University of California at Berkeley, Department of Psychology.

29.

Main

Kaplan

Cassidy

(1985). Security in Infancy, childhood, and adulthood: A move to the level of representation. In Bretherton

Waters

(Eds.), Growing Points of Attachment Theory and Research. (pp. 66–106). Monographs of the Society for Research in Child Development. No. 209, Vol. 50 nos. 1–2.

30.

Main

Solomon

(1986). Discovery of a new, insecure -disorganized/disorientated attachment pattern. In Yogman

Brazelton

(Eds.), Affective Development In Infancy, (pp. 95–124) Norwood: Ablex.

31.

Main

Solomon

(1990). Procedures for identifying infants as disorganized/disoriented during the Ainsworth Strange Situation. In Greenberg

M. T.

Cicchetti

Cummings

E. M.

(Eds.), Attachment during the preschool years: Theory, research, and intervention (pp. 121–160). Chicago: University of Chicago Press.

32.

Marvin

R. S.

(2003). Implications of attachment research for the field of family therapy. In Erdman

Caffery

(Eds.), Attachment and Family Systems: Conceptual, Empirical, and Therapeutic Relatedness (pp. 3–27). Hove, UK: Brunner-Routledge.

33.

Mayseless

(2005). Ontogeny of attachment in middle childhood: Conceptualization of normative changes. In Kerns

K. A.

Richardson

R. A.

(Eds.), Attachment in Middle Childhood (pp. 1–23). New York: Guilford Press.

34.

McCann

J. B.

James

Wilson

Dunn

(1996). Prevalence of psychiatric disorders of young people in the care system. British Medical Journal (Clinical Research Edition), 313(7071), 1529–1530.

35.

Oppenheim

(1997). The attachment doll – play interview for preschoolers. International Journal of Behavioural Development, 20(4), 681–697.

36.

Shmueli-Goetz

Target

Fonagy

Datta

(2008). The Child Attachment Interview: A psychometric study of reliability and discriminant validity. Developmental Psychology, 44, 939–956.

37.

Solomon

George

De Jong

(1995). Children classified as controlling at age six: Evidence of disorganized representational strategies and aggression at home and at school. Development and Psychopathology, 7, 447–463.

38.

Steele

(2005). The construct of coherence as an indicator of attachment security in middle childhood. In Kerns

K. A.

Richardson

R. A.

(Eds.), Attachment in Middle Childhood, (pp. 137–160). New York: Guilford Press.

39.

Target

Fonagy

Shmueli-Goetz

(2003). Attachment representations in school-age children: the development of the child attachment interview (CAI). Journal of Child Psychotherapy, 29(2), 171–186.

40.

van der Kolk

B. A.

(2005). Developmental trauma disorder: Toward a rational diagnosis for children with complex trauma histories. Psychiatric Annals, 35(5), 401–408.