Clinical Psychology and Psychotherapy Clin. Psychol. Psychother.

18, 498–511 (2011) Published online 27 May 2010 in Wiley Online Library (wileyonlinelibrary.com). DOI: 10.1002/cpp.704

Adult Attachment Insecurity and Narrative Processes in Psychotherapy: An Exploratory Study†
Sarah I. F. Daniel*
The University Clinic, Department of Psychology, University of Copenhagen, Denmark Different types of client attachment insecurity may affect the psychotherapeutic process in distinct ways. This exploratory study compared the in-session discourse of clients with dismissing and preoccupied attachment states of mind on Adult Attachment Interviews conducted prior to therapy in the context of a randomized clinical trial of psychoanalytic and cognitive–behavioural psychotherapy for bulimia nervosa. In a subsample of six sessions from each of eight therapy dyads, preoccupied clients were found to talk more and have longer speaking turns than dismissing clients, who in turn generated more pauses. Using the Narrative Processes Coding System, preoccupied clients were found to show more narrative initiative; whereas, differences in terms of narrative process modes were not as clearly interpretable. Contrary to expectations, the two insecure states of mind were equally different in the relationship-focused psychoanalytic therapy and in the symptom-focused cognitive–behavioural therapy. Suggestions for further investigations of the in-session discourse of clients with different attachment states of mind are given. Copyright © 2010 John Wiley & Sons, Ltd. Key Practitioner Message: • Client adult attachment patterns are likely to affect psychotherapeutic processes. • Client insecure dismissing and insecure preoccupied attachment may be detectable in distinct in-session discourse styles. • More research is needed on in-session manifestations of client attachment patterns and their relation to therapeutic outcome. Keywords: Adult Attachment, Psychotherapeutic Processes, Narrative Process, Verbal Productivity

INTRODUCTION
In the literature on client characteristics that impact psychotherapy process and outcome, a growing attention is given to the topic of adult attachment (Daniel, 2006; Norcross, 2002; Obegi & Berant, 2009). Attachment theory was originated by Bowlby (1969/1997, 1973/1998, 1980/1998) and deals with the central role that early relations to caregivers play in psychological health and in shaping the ways in which later interpersonal relationships are approached and handled. Adult attachment patterns or styles are developmentally formed ways of

* Correspondence to: Sarah I. F. Daniel, Department of Psychology, 2A Oester Farimagsgade, 1353 Copenhagen, Denmark. E-mail: sarah@selskabet.org † This manuscript was prepared as part of a PhD project at the University of Copenhagen funded by a grant from the Danish Research Council for the Humanities. Transcription of therapy sessions was generously supported by the Aase & Ejnar Danielsen Foundation.

regulating affect and processing information with respect to close interpersonal relationships (Mikulincer & Shaver, 2007). Since adult attachment patterns are related to attitudes towards seeking and receiving help and to distinct ways of engaging in emotionally significant relationships, client attachment patterns are likely to affect psychotherapeutic processes. One of the central approaches to measuring adult attachment patterns is the Adult Attachment Interview (AAI) (George, Kaplan, & Main, 1996; Main, Goldwyn, & Hesse, 2002), which has been used extensively in developmental psychology research and has demonstrated good psychometric properties (Bakermans Kranenburg & Van Ijzendoorn, 1993; Crowell, et al., 1996; Sagi, van IJzendoorn, Scharf, & Koren-Karie, 1994). The AAI operationalizes adult ‘states of mind with respect to attachment’ in terms of different patterns of coherence and collaboration in narrating about attachment-related experiences (Main, Hesse, & Goldwyn, 2008). The classification system distinguishes between three main categories: secure interviews are balanced, open and coherent;

Copyright © 2010 John Wiley & Sons, Ltd.

although it is more common in many clinical groups (Dozier. A few studies have inquired into the in-treatment behaviour of clients with different attachment states of mind. 1999. Teti et al. Dismissing mothers were less emotionally committed to treatment and more likely to prefer simple Copyright © 2010 John Wiley & Sons. positive evaluations of childhood experiences and lack of convincing autobiographical evidence. Stovall. persons with preoccupied states of mind are continually caught up in attachment-related experiences and hyperactivate the emotional responses related to the attachment behavioural system.e. Persons with secure states of mind shift their attention flexibly between attachment and non-attachment related domains. whereas. 1999). the contrast between dismissing and preoccupied attachment insecurity merits attention. However. at the same time. there is a fair amount of evidence that attachment states of mind affect how clients engage in treatment and that secure clients are generally more involved and committed to treatment. & Cicchetti. and often marked by inherent contradictions between general. more dismissing) were less likely to seek out help... Heinicke and Levine (2008) found that secure mothers were more and increasingly involved in home visitation work. van Ijzendoorn. preoccupied clients may benefit more from working with therapists who do not amplify their emotional experiences. and a greater decrease in avoidance of contact with staff in the course of 10 months of treatment. & Albus. 2007). Both studies had insufficient numbers of preoccupied mothers to consider that group. In contrast. Levy et al. Psychol. Schuengel. and that clients who were high on deactivation (i. Dozier and colleagues investigated correlates of attachment in a psychiatric setting using Kobak’s Q-sort method of coding AAIs. Dozier. insecure mothers were less and decreasingly involved. whereas. more research is clearly necessary to posit such differential treatment implications confidently. Evidence is emerging that it may be possible to facilitate client movement in the direction of secure states of mind through psychotherapy (Fonagy et al. Such differences in attention. Few studies have been able to compare the two primary groups of insecure clients. The present study constitutes an explorative investigation of differences between psychotherapeutic process with dismissing and with preoccupied clients. Rogosch. Dozier (1990) found that greater security was associated with more compliance with treatment. Zegers. more preoccupied) clients.e. partly because the preoccupied category is relatively infrequent in most western samples. 499 companionship to supportive therapy or problem-solving help than secure mothers were. 2000. Psychother. Tyrrell. 18. 2008). studying mothers who participated in a relation-based early intervention programme. 2008). Adam. such research may also throw more light on the mechanisms behind possible differential treatment implications. Fonagy et al. Teague. 1995. Ogawa. whereas. 1996. and Janssens (2006) found that secure attachment was related to a greater increase in reliance on the primary contact figure in the institution. Slade. since the challenges of working with clients who appear distant and emotionally unengaged are likely to be fundamentally different from the challenges of working with clients who appear overinvolved and demanding. In a related study. Persons with dismissing states of mind direct their attention away from attachment-related experiences and from the attachment component of relationships in an attempt to deactivate the attachment behavioural system and related feelings of needing or longing for comfort. and Egeland (1997) investigated client behaviour in an intervention programme for mothers at risk. 1999. emotion-regulation and narrative coherence are clearly of relevance to psychotherapy. In contrast Clin. Toth. 1995. Given the differences between the dismissing and the preoccupied categories. Main. Thus. Korfmacher. and Fallot (1999) suggests that dismissing clients may benefit more from working with therapists who challenge their distancing from emotions and interpersonal closeness. more likely to reject help and less likely to self-disclose as compared with more hyperactivating (i. A study by Tyrrell. Ltd. (2008) studied mothers’ commitment to an early intervention programme for at-risk mothers with premature babies and found that secure mothers were more involved compared with dismissing mothers.Attachment and Narrative Process in Therapy dismissing interviews are overly brief and unemotional. 2009). it is likely that the ‘road’ to security differs for these clients. remain psychologically separate. which requires clients to engage in a collaborative dialogue about emotionally difficult topics and to provide a reasonably clear account of their situation in order for the therapist to empathize and intervene with precision (Daniel.. and Lee (2001) found that more deactivating clients spent less time on task when instructed to discuss an area of interpersonal difficulty with their case managers. In a study of adolescents treated at a youth institution. They are able to recognize their need for attachment figures and. Mikulincer & Shaver. 2006.. Dozier. These different narrative ‘surface structures’ are thought to be related to particular ways of regulating attention and emotion with regards to attachment that have their roots in childhood experiences (Magai. However.. and this may have important implications for psychotherapeutic technique (Holmes. A first step in that direction is more research on the in-treatment manifestations of the different insecure attachment patterns. Also. preoccupied interviews contain long-winded and confusing accounts that are marked by continued emotional entanglement. 1999. Lomax. van Ijzendoorn & Bakermans Kranenburg. In addition to providing information about the degree to which client attachment patterns may be detectable by clinicians through characteristic in-session processes. 498–511 (2011) .

Fonagy et al. there are no published studies systematically investigating the in-session discourse of psychotherapy clients. Because of the limited knowledge in the area. the number of words uttered). secure transcripts were in between and dismissing transcripts were the lowest. in that the process of seeking help might activate the attachment behavioural system and thus. PPT and CBT for bulimia are very distinct treatment forms that make it possible to explore whether intreatment expression of discourse patterns related to attachment states of mind is conditional on therapeutic exploration of attachment relationships. The NPCS conceptualizes the therapy dialogue as a narrative process in which client and therapist collaboratively explore and re-evaluate aspects of client ‘life stories’ (Angus. this study also investigated the narrative material generated using the Narrative Processes Coding System (NPCS) devised by Angus. this study investigated three variables related to this: (a) the total amount of verbal production by the client in sessions (i. In contrast.. since theoretically. It is an integrative psychotherapy process coding system. 2007). Thus. Ramacciotti et al. & Hesse. whether they ‘hold the floor’ for long segments or quickly return the speaking turn to therapists. 18. Daniel The current study was conducted in the context of a randomized clinical trial of psychoanalytic psychotherapy (PPT) and cognitive-behavioural psychotherapy (CBT) for bulimia nervosa. who consistently pushes for immersion in attachment themes. To date.. In addition to the variables related to verbal productivity. These variables speak to how active clients are in contributing to the therapy dialogue. F. attachment issues are seldom discussed in detail. Ward et al. Dismissing interviews cut short the examination of attachment relationships and offer little detail. Tesman. Main et al. 2004. 1999).. this is not necessarily the case. Lewin. Dismissing and preoccupied AAIs differ markedly in their narrative style. and the verbal activity of the therapist may be very different from that of the AAI interviewer. Beebe. there was no overlap of classification in persons assessed with both interviews. and whether they pause before responding or fall silent in the course of their speaking turns. Parish & Eagle. the related discourse patterns (Dozier & Bates. Wallin. Goldwyn. Ltd. Thus. 1989). perhaps. it may be hypothesized that attachment-related differences in in-session discourse of clients will be more easily detectable in PPT than in CBT. which is a measure of narrative concreteness. Feldstein. (1996) found that when applying an AAIlike classification strategy to an interview about employment experiences. Crown. 498–511 (2011) . which play a central role in the AAI coding system (Main.e. the present study took an exploratory and descriptive approach to the investigation of in-session discourse. 2003). CBT is highly therapist-directed. and both dismissing and preoccupied patterns are represented in this group (Candelori & Ciocca. Thus. Psychother. This sample is relevant to the present purpose in two ways. Crowell et al. the discourse characteristics exhibited in the AAI are thought to be specifically related to the activation of the attachment behavioural system. and (c) the number of pauses occurring in sessions. Malatesta. Jaffe. 1999. (b) the length of client speaking turns. 2001. insecure attachment states of mind are prominent in bulimia nervosa and other eating disorders. 1998. Mergenthaler & Bucci. Hardtke. specificity and imagery devised by Bucci (1995. Psychol. S. Preoccupied interviews offer extended and often long accounts with many (frequently irrelevant) details. 1991. & Jasnow. the focus of the present study is on dialogical and narrative features.. Furthermore. which is brought about by the interviewer’s query into attachment history (Main. 1996. they may be flooded by emotion to the point where cooperation with the interviewer breaks down. 2008). Bouffard. in terms of several discourse dimensions. They evaluated the transcripts with respect to emotion words. On all these measures. Clin. the prototypical dismissing and preoccupied interviews represent opposite extremes with the secure category occupying the ‘middle range’. and Levitt (1996). 1995). In the literature on client attachment and psychotherapy. Thus.500 to the primary focus on engagement and commitment in previous studies. In fact. a tendency that is also found in other parts of the attachmentrelated literature (Isabella & Belsky. On the other hand. PPT for bulimia is based on the theoretical premise that the eating disorder symptoms have their roots in difficulties in affect regulation that are primarily interpersonal in their origin. 2001. attachment relationships are often discussed in PPT. 2001). It might be that psychotherapy is an attachment-relevant situation by design. it is frequently assumed that the narrative characteristics exhibited in the AAI are similarly evident in psychotherapy process (Fish & Dudas. However. whose attachment states of mind have been independently assessed. Buchheim and Mergenthaler (2000) conducted a computer-assisted linguistic analysis of 30 AAIs categorized into the three main attachment states of mind. Holmes. CBT for bulimia is a highly structured treatment that focuses on the eating disorder symptoms and takes a behavioural approach towards regulating these. I. in psychotherapy. Since one of the features that typically distinguishes dismissing and preoccupied speakers in the AAI is verbal productivity. & Rotondi-Trevisan. Culver. First. 2002. they are unemotional or attempt to minimize emotion. PPT is non-directive. preoccupied transcripts scored highest. 2001. whereas. Consequently.. abstraction and referential activity. this has not been demonstrated empirically. Copyright © 2010 John Wiley & Sons. 2004). & Shepard. a demonstration that clients’ in-session discourse is parallel in style to their discourse in the AAI would not be a trivial finding. Sandros. Indeed. 2006. attachment relationships are not necessarily discussed. Second. at present.

Compared with dismissing clients. Clients are recruited through advertisements in magazines. They will say more. schizophrenic spectrum disorders. 2002). the three-way distribution of attachment categories was 21. (d) somatic illness that will interfere with psychotherapy. Jurist. 1996. 18. Psychol. The therapy dialogue is divided into narrative sequences according to shifts between three different narrative process modes: (a) external narrative process where life events are related or facts or practical issues are described. & Target. 3. which can be interpreted as an index of how active each speaker is in ‘shaping’ the pattern of shifting between topics and narrative modes. METHOD Study Context The study was carried out in the context of a randomized clinical trial of PPT and CBT for bulimia nervosa (N = 70). Clients are required to (a) be at least 18 years old. Main et al. (b) fulfil the Diagnostic and Statistical Manual of Mental Disorders. dismissing clients will tend to focus on the ‘landscape of action’. (b) internal narrative process where feelings and other subjective experiences are articulated. 1990). but in an earlier study with a similar client population. manic and bipolar disturbances. 1993. 2002). Thus. or somatoform disorders if the disorder is judged to be a major obstacle to treatment of bulimia. the internal narrative mode. The sample consisted of four dismissing clients (two in PPT. in public places and on the Internet. (e) pregnancy or other plans interfering with treatment. would theoretically be expected to be less common with dismissing clients than with preoccupied clients. and preoccupied clients will tend to focus on the ‘landscape of consciousness’. Also. which consists of descriptions of feeling states. clients and therapists complete a brief session evaluation form. two in CBT) and four preoccupied clients (two in PPT. dissociative disorders. The following hypotheses were posited: 1. 1999). as argued by Angus et al. or they might not appear at all in CBT. Finally.Attachment and Narrative Process in Therapy which can be used in coding both of the relatively different therapy forms included in the sample. it is tentatively hypothesized that. two Clin. 2006). Participants Eight therapies from the Bulimia Trial were selected based on client intake AAI category and treatment type. whereas. Ltd.. psychological reality. Clients are randomly assigned to the two treatments with block randomization stratifying for current psychopharmacological treatment and personality disorder. Dismissing and preoccupied clients may differ in their attention to and willingness to discuss outer and inner/psychological realities with their therapists. The differences between preoccupied and dismissing clients will be larger in PPT than in CBT. 2004).1% preoccupied (Sandros. The attachment data for the entire client sample are still not available. 2. Preoccupied clients will engage in more internal and reflexive narrative Copyright © 2010 John Wiley & Sons. (c) current alcohol or substance abuse. dismissing clients will engage in more external narrative processes. Gergely. whereas both dismissing and preoccupied clients will have difficulties in coherently integrating these two narrative landscapes (Fonagy. outer reality. After each session. preoccupied clients may be expected to engage more in the therapeutic dialogue. the three narrative modes can be conceptualized in terms of the ‘dual landscape’ of narrative outlined by Bruner (1986. The internal and reflexive modes comprise the ‘landscape of consciousness’ or the ‘subjective’. In summary. all clients complete a battery of questionnaires assessing various aspects of personality and psychopathology and participate in diagnostic interviews and in the AAI (George et al. 36. (2004). Psychother. Preoccupied clients will show more narrative initiative than dismissing clients. or moderate or severe depression. This classification is purely descriptive and does not imply an evaluation of the quality of these processes. Exclusion criteria are (a) concurrent organic disorders. which may be discernible in a higher degree of narrative initiative. Therapy sessions are recorded on minidisc by the therapists. the aim of the current study was to explore possible differences between in-session discourse with dismissing and preoccupied clients and to generate further suggestions for studies in the area. and (f) difficulties in speaking and/or understanding the language.8% dismissing and 42. and (c) give informed written consent to participate in the study. Fourth Edition (DSM-IV) criteria for a diagnosis of bulimia nervosa. substance-related disorders. At intake. 498–511 (2011) . which is presently conducted at a university-based psychotherapy research clinic (Lunn & Poulsen. and through a network of clinics and psychiatric hospitals. 501 processes. Levitt & Angus. The external narrative mode comprises the ‘landscape of action’ or the ‘objective’. Preoccupied clients will be more verbally productive in sessions than dismissing clients.0% secure. The NPCS has been found to distinguish between different therapy approaches and to make meaningful distinctions between different types of process in in-session dialogue (Levitt. Each narrative sequence is also coded with respect to who initiates it. hold the conversational turn longer and will pause less. and (c) reflexive narrative process where events or feelings are analysed and discussed with respect to patterns or meaning. (b) concurrent anxiety disorders..

and their responses aim at helping clients to acknowledge and Copyright © 2010 John Wiley & Sons. clients who dropped out before the completion of 20 therapy sessions were not included. The clients included in the study were the first eight clients in the Bulimia Trial. only the three-way classification of secure. The therapy consists of weekly sessions for a period of 2 years.38.0. who fulfilled the selection criteria in terms of attachment and distribution across therapy types. The interviewee is asked about the relationship to childhood attachment figures and attachment-related topics such as separations. the unresolved classification may be assigned based on brief passages in interviews that are otherwise secure. It is a highly structured therapy consisting of 20 sessions with fixed agendas.. 2002). dismissing or preoccupied in their narrative style. and to be able to sample from a comparable range of sessions. two out of three additional coders agreed with the classification made by the author.91. 2008. Daniel reflect on their inner world of experience. Six therapists were female and two were male. No clients in the sample had the same therapist. and therapists consistently focus on the eating disorder. who is certified as an AAI coder. F. Thus. scored the AAIs from the intake assessment. and for whom usable sound recordings of the therapy sessions were available. preoccupied: M = 1.. Measures AAI The AAI is a semi-structured interview with 20 questions and standardized probes (George et al. 1996).29. preoccupied: M = 26.502 in CBT). & First. Three other certified AAI coders scored a random sample of four of the eight AAIs to check inter-rater reliability. The therapists were trained according to the specific treatment manuals (Fairburn. Based on scores on the ‘state of mind’ scales and on the presence of global category descriptors. Therapists allow clients to direct the choice of topics.. Crowell et al. 1993. The AAI has been found to have good reliability and discriminant validity (Bakermans et al. Psychother. I. 1990). Fairburn. four were students and four were employed. SD = 7. 1996). . To code the AAI. dismissing or preoccupied was used. SD = 0. 2005) and were supervised bi-weekly by Susanne Lunn and Stig Poulsen (PPT) or Christopher Fairburn (through video conference) (CBT). the AAI was administered by post-graduate psychology candidates who were trained and supervised by certified AAI coders. All had completed the equivalent of high school education.. The author. dismissing even if they were also unresolved. Five of these scales evaluate aspects of childhood experience as inferred by the coder. In contrast to the three primary AAI classifications. the author was blind with regards to which treatment type the clients had been assigned to. Lunn & Poulsen. and the therapists were not informed about client attachment patterns. In two cases. coders have to attend a 2-week training workshop with approved trainers and to achieve at least 80% agreement with the norm on a subsequent reliability check consisting of 30 interviews (Hesse. unresolved or cannot classify. & Kraemer. preoccupied. interviews categorized as Clin.5 years (standard deviation [SD] = 5. all three additional coders agreed with the classification made by the author. Walsh. Williams.24) (possible scores range from 0 to 4). Clients are given homework assignments. 498–511 (2011) Treatments PPT for Bulimia PPT for bulimia (Lunn & Poulsen. etc. Interviews classified as unresolved or cannot classify are also given a secondary classification as secure. When selecting the current sample. interviews are assigned to one of five major categories: secure. The clients were women of Scandinavian origin with an average age of 25. 1999). SD = 0..3). CBT for Bulimia CBT for bulimia was developed by Christopher Fairburn (2008) and is one of the best validated treatments for bulimia (Agras. 2000). the bulimic symptoms are not necessarily discussed in every session. Psychol. The therapists were experienced psychologists or psychiatrists with expert training in either PPT or CBT. and in two cases. To keep the sample as homogenous as possible except for type of attachment insecurity. being ill. dismissing and/or preoccupied (Main et al. 1994) (dismissing: M = 1. dismissing. In general.3) and in terms of overall psychological distress as measured by the general severity index of the Symptom-Checklist-90-R (Derogatis. clients were included as e. Two clients in each group were diagnosed with one or more personality disorders based on the Structured Clinical Interview for DSM-III-R Personality Disorders (Spitzer. The groups of dismissing and preoccupied clients were relatively comparable in terms of age (dismissing: M = 25.. Gibbon. The other 11 scales evaluate different aspects of the speaker’s ‘state of mind with respect to attachment’. The aim of CBT is to break the vicious cycle of dieting and bingeing by establishing a regular eating pattern and correcting the cognitive distortions related to the disorder. that is. 2005) is a nondirective therapy based on a psychoanalytic understanding of the bulimic symptoms that are thought to be rooted in difficulties in affect awareness and regulation. S. SD = 3. Ltd.g. and these are systematically explored.2. In the Bulimia Trial. losses. 18. Wilson.0. When scoring the AAIs. The interview is transcribed verbatim and rated on 16 nine-point scales.

in which sessions from therapies with bulimic clients were transcribed and coded with NPCS by the author. On recommendation from Lynne Angus. and (4) the degree to which they had trust in their therapist in the session. During the training period. who was trained and supervised by Lynne Angus. Pauses were timed by the transcriber and were indicated in the session transcripts. Training of NPCS Coders Eleven university students volunteered to code the therapy sessions. the coders received detailed feedback. 2006). The coders practiced on material from a pilot study.86) (Angus. Kappa was calculated based on the number of topic shifts agreed upon. Counts of Words. the number of shifts identified only by rater B and the number of agreements on the non-presence of Clin. Sessions are first subdivided into topic segments according shifts in semantic content. and difficulties in rating the sessions were thoroughly discussed. who coded one or two full sessions between meetings. 503 based on client ratings of session difficulty on the SEF-C. All 48 sessions were transcribed by a professional agency using a slightly modified version of the transcription guidelines developed by Mergenthaler and Stinson (1992). 498–511 (2011) . Client Version (SEF-C) The SEF-C is a four-item questionnaire developed specifically for use in the Bulimia Trial. 1996). The topic segments are further subdivided in terms of three narrative process modes: external. & Angus. Scores are summed to give a total score that may range from 0 to 63 (Beck.Attachment and Narrative Process in Therapy unresolved are most frequently secondarily classified as preoccupied. 18. yielding 991 transcript pages. 2003). In the Bulimia Trial. respectively. Speaking Turns and Pauses Counts of the number of words uttered by clients per session and per speaking turn were conducted using the C-I-SAID software (Cartwright. Topic segments and narrative sequences are also coded with respect to who initiates them. They were split in two teams that specialized in dividing the sessions into topic segments and in dividing topic segments into narrative sequences. (2) the degree to which they consider the session to have been difficult. The BDI-II is widely used and has shown good psychometric properties. Beck Depression Inventory II (BDI-II) The BDI-II is a 21-item self-report instrument that asks clients to rate different depressive symptoms on a fourpoint scale.. NPCS inter-rater reliability in terms of Cohen’s κ was calculated using methods described in Bakeman and Gottman (1997). each coder coded at least seven full sessions. good inter-rater agreement has been found on identification of topic segments as well as on coding of narrative process (Cohen’s κ = 0. Sessions without a usable sound recording and sessions with participation of family members (in CBT) were excluded.78–0. Session Evaluation Form. The intake session that consisted mostly of history-taking and introduction to the treatment was not included. Greenberg. the three sessions experienced as most difficult and the three sessions experienced as least difficult were selected. Three seconds was chosen as the lower limit of meaningful silence based on previous research (Frankel. Murray. NPCS The NPCS is a psychotherapy process coding system that is applied to session transcripts (Angus et al. Steer. and two of the preoccupied clients in the current sample were also unresolved. To obtain a varied sample of therapy process. The number of pauses at or above 3 seconds in all sessions were counted using the search and replace function of an ordinary word processor. & Brown. All coders attended an introductory lecture on the NPCS. the number of shifts identified only by rater A. For each client. & Hardtke. 1996). Ltd. The author had five to six 2-hour meetings with the coders. (3) the degree to which they felt understood by their therapist in the session. Line numbers in the therapy transcripts served as a rough ‘time unit’ in comparing the coding of raters. sessions were selected Copyright © 2010 John Wiley & Sons. Whereas client ratings of items 1. External process refers to the description of events or facts. After each session. and reflexive process refers to analysis and interpretation of events or feelings. Clients categorized as cannot classify were not included. 3 and 4 are generally highly skewed towards the high end of the scale. ratings of session difficulty utilize the whole seven-point scale. internal or reflexive. At the meetings. clients are requested to give their ratings on a seven-point Likert scale of (1) the degree to which they consider the session to have been valuable/useful. internal process refers to the description of feelings or other inner experiences. Psychother. Six were undergraduate psychology students. coders were considered to agree on a topic shift if they placed shifts less than six transcript lines apart (the average length of topic segments in the material was 50 transcript lines). With raters trained for 20–25 hours. it was administered at intake as part of a larger battery of self-report instruments NPCS Inter-rater Reliability Procedure Selection of Therapy Sessions A sample of six of the first 20 sessions from each of the eight therapy dyads was selected for further analysis. 1999). Psychol. and five were graduate psychology students. Levitt. Levitt.

75–0. Since the material was considered difficult to code.71 (range 0. and the narrative sequencing coders achieved an average agreement with the author of κ = 0. it is possible to take into consideration the random variation between superordinate clusters of data (in this case. Since sessions from the same dyad are non-independent.85). the coders worked in pairs. Since some of the dependent variables might change with time across the six months of therapy. Because of the low inter-rater agreement on the internal narrative mode.5%) were coded as internal by one or both coders. The main explanatory variables: client attachment state of mind and therapy type were measured at the level of therapy dyads. therapy dyads). which occurred only in CBT. . The topic segmenting coders achieved an average agreement with the author of κ = 0.90.57–0. Only the consensus coding was used in the further analyses. the coders negotiated a consensus coding. A small number of narrative sequences (21 out of 1292) could not be coded Copyright © 2010 John Wiley & Sons.94). inter-rater agreement on narrative initiative was calculated based on all instances where coders started new narrative sequences within two transcript lines of each other.78 (range 0. internal or reflexive. and thus raw counts of the number of words and pauses Clin. The dependent variables examined were all measured at the level of sessions. In distinguishing external from other codes.75 (range 0. The narrative sequencing coders achieved an agreement of κ = 0. in distinguishing internal from other codes. There was no evidence of formation of coding subcultures in the coder pairs. Even though most sessions lasted about 50 minutes. since they were considered ambiguous with respect to whether they should be allocated to the ‘landscape of action’ or the ‘landscape of consciousness’.90). Inter-rater reliability on narrative sequencing was calculated as time-based kappa’s of the line-by-line agreement on narrative mode categorization using the GSW software (Bakeman & Quera. F. At the end of training. NPCS Coding Procedure The NPCS coders were blind with regards to client attachment categorization as well as the hypotheses of the study. 18. the coders achieved an agreement of κ = 0. it became necessary to include an extra coder in the team of topic segmenters. The coder. only 148 lines (0. The agreement on coding of narrative initiator was κ = 0. Each coder coded between 12 and 24 sessions. Since coding of who initiates narrative sequences can only be compared when coders agree on the delineation of narrative sequences. the total number of transcript lines minus the number of shifts agreed upon and the number of shifts disagreed upon). and direct instruction from the therapist with respect to what clients should do (n = 17). since the intra-class correlation coefficients of the dependent variables range from 0. Psychother. Out of the 30 088 transcript lines in the data material.20. S. In the present study.59 to 0. I. and in distinguishing reflexive from other codes.504 shifts (i. Subsequently. Towards the end of the coding period that lasted about a year.88 (range 0. These were descriptions of events taking place in dreams (n = 4). was similarly trained by the author and achieved a reliability of κ = 0. During the coding period.. All analyses allowed for random intercepts for therapy dyads. and when they had negotiated a consensus coding. Ltd.77) on classifying transcript lines as external.83–0. such an approach is definitely called for.83. the agreement was κ = 0. Each of the 48 sessions was first given to two topic segmenting coders.81 (range 0. the topic segmenting coders achieved an agreement of κ = 0. who had an MA in psychology. the marked transcript was passed on to two narrative sequencing coders. the agreement was κ = 0.72.83).8–73 minutes). Cohen’s κ was also calculated separately for each of the three narrative process modes. data analysis was conducted by means of multilevel mixedeffects modelling (Rabe-Hesketh & Skrondal. these sequences were excluded from the analysis pertaining to narrative process modes.79) on the line-by-line presence of topic shifts.71. 2008). and since there might also be dyad-dependent variation in the rate of change. RESULTS Data Set and Data Analysis The data set consisted of six sessions from each of eight therapy dyads. given that all sessions were subsequently consensus-coded. which occurred only in PPT. the sessions were of unequal duration (range 36. 498–511 (2011) Inter-rater Agreement and Stability of NPCS Coders When coding the study material. 2007). which is not accounted for by the explanatory variables: client attachment state of mind and therapy type. This level of agreement was considered satisfactory.68 (range 0. the analyses also allowed for a random slope for time (defined by session number) within therapy dyads. These sequences were also excluded from the analysis pertaining to narrative process modes. In multilevel analysis. the topic segmenting coders achieved an average agreement with the author of κ = 0.72–0. Daniel for narrative process mode. Psychol.59–0.91. distributed as evenly across therapy types and client attachment categories as allowed by coding logistics.66–0.e. This indicated problems with the coding for internal process. Each member of a pair first coded each session individually and submitted the coding to the author who calculated inter-rater agreement. and the narrative sequencing coders achieved an average agreement with the author of κ = 0. all coders were asked to recode a session from the training material to check for possible rater drift. which was also very rare in the material.

& Brown. AAI unresolved score and session difficulty were also considered. possible main effects of client attachment state of Copyright © 2010 John Wiley & Sons. p = 0.4 words per session minute.0.0. the difficulty ratings of dismissing clients tended to be higher than those of preoccupied clients (preoccupied: M = 3. Howard.001).3. which has in turn been related to the dismissing category in some studies (Patrick. Client depression. z = 2.0.2.3. though not significantly.38. standard error [SE] = 10. neither alone or in interaction with client attachment or with time.91. SD = 2. p < 0. z = −2. Psychol. 18.027) as well as the main effects of client attachment state of mind (B = −42. For the number of client words per session minute. z = 3. SE = 10. A limitation of the current sample was that two of the four preoccupied clients were also categorized as unresolved on the AAI. preoccupied clients uttered 84. To control for this difference. the number of client words in sessions was divided by session duration in terms of minutes. 1997). SE = 1.75.7. In the present sample. Client depression did not affect this variable. z = −3. whereas. z = −2.0. SD = 8. client-rated session difficulty was included in the analyses as a fixed effect. Castle. The proportion variables were subjected to a logit transformation for inclusion in the linear analyses. z = 3. higher in the group of dismissing clients (M = 20. SE = 1.06. 505 mind.2. SE = 24. as session difficulty increased.2 words per speaking turn. whereas. Although not statistically significant.21.13.001). preoccupied clients in PPT uttered 138. In CBT. with session difficulty set to 4. The number of words per client speaking turn decreased as session difficulty increased and as client unresolved score increased. z = −2.001) and a significant main effect of session difficulty (B = −3. SD = 1. Based on this model. SE = 12. p < 0.40. a significant main effect of therapy type (B = −48.0). starting with the full model and gradually excluding non-significant terms. In CBT. Lowered speech rate has been related to depression (Ellgring & Scherer. there was no evidence of time-effects and no significant interaction between client attachment and therapy type. To control for the possible effect of this difference.Attachment and Narrative Process in Therapy were not directly comparable. SD = 1.0. The number of client words per session minute decreased. preoccupied clients in PPT uttered 132.6 words per session minute.002). The final model included a significant Client Attachment × Time interaction (B = −2. p = 0.. There was a significant main effect of client attachment state of mind (B = 31.027). SE = 15.1.080) and of time (B = 4.8.96. 498–511 (2011) Verbal Productivity Variables When analyzing the variables related to verbal productivity. z = −1. z = −3.001). To maximize session variability. dismissing clients were estimated to utter 46.97. In the present sample. and since session difficulty might conceivably affect some of the variables studied. the preoccupied clients had higher unresolved scores than the dismissing clients (preoccupied: M = 4. dismissing: M = 1. which can range from 1. SD = 8.0. i.1 words per speaking turn. The NPCS variables were evaluated in terms of proportions. z = −5.25). therapy type and time.5. p = 0.004) and a significant main effect of client unresolved score (B = −9. since the current level of depression of clients might affect the verbal productivity variables. For the number of pauses per session hour. it was estimated that dismissing clients in PPT uttered 107. and all possible two-way interaction terms were considered.18. Client depression and client unresolved score were not related to this variable. p < 0. & Maughan.26. client BDI score was included in these analyses as a fixed effect. Psychother. Ltd. depression scores on the BDI-II (Beck. this study instead evaluated relative durations of narrative modes in sessions (in terms of word counts) to more accurately reflect the ‘time-budget’ representation of the different narrative process modes (Bakeman & Gottman.41. 1994). in the present sample. and the number of pauses was divided by session duration in terms of hours to yield comparable and interpretable indices. therapy type had no significant effect.8. a significant main effect of session difficulty (B = −3. There was a significant main effect of client attachment state of mind (B = 61.55. 1996) ranged from 8 to 29 and were slightly. SE = 1. a significant main effect of therapy type (B = −61.2. p = 0.7 words per speaking turn. whereas. whereas.52. there was no time-effect and no significant interaction between client attachment and therapy type. Steer.6 words per session minute. SE = 3. sessions had been chosen based on high or low ratings of session difficulty on the SEF-C.5.4 words per session minute.e. preoccupied clients uttered 77. p = 0.37. Main effects of client BDI-II score.84.8) than in the group of preoccupied clients (M = 17. client unresolved score was included in all analyses. a score of at least 5. Based on this model. it was estimated that dismissing clients in PPT uttered 70. Figure 1 shows the estimated number of pauses per session hour with dismissing and preoccupied clients Clin.88. client unresolved score and session difficulty did not affect this variable. For the average number of words per client speaking turn. dismissing clients were estimated to utter 22. SD = 2. p = 0.0 to 9.016). Although narrative process modes have usually been evaluated in terms of relative frequencies of sequences categorized into different modes. 1996).84.7 words per speaking turn. p = 0.41.50. However.46. Hobson. . dismissing: M = 4.0 has to be given on a rating scale for unresolved loss or trauma.88. Consequently.1.03). SE = 0. proportion of narrative sequences initiated by the client (versus the therapist) and proportion of words coded as external (versus reflexive). with session difficulty set to 4 and client unresolved score set to 1. To be assigned to the category unresolved.

Psychother. p < 0.8%. Narrative Initiative and Narrative Process Modes When analysing the NPCS variables.001) and a significant main effect of therapy type (B = −2. the corresponding number for preoccupied clients was 81. in PPT. p = 0.10.001). 498–511 (2011) .16.9 with dismissing clients. 18.78.014). whereas. Client unresolved score and session difficulty did not affect this variable. Client unresolved score and session difficulty did not affect this variable. Psychol.002).067). whereas. z = −3. I.2 with preoccupied clients. z = −2.0%. In CBT.87.506 140 Dismissing 120 Pauses per session hour 100 80 60 40 20 0 Preoccupied S. SE = 0. F. with preoccupied clients in PPT it is 31.61. Daniel 2 4 6 8 10 12 Session 14 16 18 20 Figure 1.0%. the number of pauses in therapies with preoccupied clients tends to remain stable across the 20 sessions. For the proportion of narrative sequences initiated by clients (versus therapists) in each session there was no time-effect and no significant interaction between client attachment and therapy type. although the difference here is much smaller.70.23.05. and all possible two-way interaction terms were considered.04. with dismissing clients in CBT it is 68. p < 0. There was a significant main effect of client attachment state of mind (B = 0.05. it was 20. of therapy type (B = 4. At session 10. Based on this model. and time.51. z = 1. and the number of pauses tends to increase across the 20 sessions. possible main effects of client attachment state of mind. both Client Attachment × Therapy Type and Therapy Type × Time were statistically significant. Based on this model. SE = 1. z = 4. therapy type.6%.2%. the estimated percentage of narrative sequences initiated by dismissing clients was 68.46.4%.001) and of time (B = 0. Ltd. Figure 2 shows the estimated percentage of narrative process categorized as external (versus reflexive) with dismissing and preoccupied clients in each therapy form across the 20 sessions based on this model. This corresponds well to an initial strong behavioural focus in CBT. Clin. the corresponding number for preoccupied clients was 32. SE = 0. p = 0. the proportion of external process tends to decrease across the 20 sessions.17.46. the two therapy types are more different in the early sessions and more alike around session 20. as well as the main effects of client attachment state of mind (B = 1. For the proportion of narrative process categorized as external (versus reflexive) in each session. p = 0. z = 4. In CBT. it is the other way around. z = −12. p = 0. whereas. SE = 0. the proportion of external process is higher with dismissing clients than with preoccupied clients. in PPT there is a slight increase in the proportion of external process. p < 0. and with preoccupied clients in CBT it is 39. the estimated proportion of external process with dismissing clients in PPT is 25. In CBT. SE = 0. evidenced in a high proportion of external process. z = 2. in PPT.07.6%. which eventually gives way to an increased focus on the cognitive distortions behind the eating disorder. SE = 0. the estimated percentage of narrative sequences initiated by dismissing clients was 19. Copyright © 2010 John Wiley & Sons.3%. Main effects of client AAI unresolved score and session difficulty were also considered. Thus. In therapies with dismissing clients. in terms of proportions of narrative modes. starting with the full model and gradually excluding non-significant terms.98.014). Linear prediction of number of pauses ≥ 3 seconds per hour across sessions for each client attachment state of mind across the 20 sessions based on this model. The final model included Client Attachment × Therapy Type (B = −1.12. SE = 0. resulting in higher proportions of reflexive process. whereas. Therapy Type × Time (B = −0. whereas. there are more pauses.17. whereas. the estimated number of pauses per hour at session 10 was 91.83.71.27.

and by client–therapist relationship factors. Thus. Also. and the small sample size of this study increases the risk that accidental combinations of particular client and therapist characteristics in the eight dyads may have influenced the results. future studies should consider secure and unresolved attachment as well. consisting only of women who completed at least 20 sessions of outpatient psychotherapy for bulimia nervosa. The results should be interpreted with caution. Ltd. 18. an attempt should be made to relate the verbal productivity variables to treatment outcome. Thus.CBT 507 Figure 2. 498–511 (2011) . by other therapist variables than therapy type. Assuming that the difference in verbal productivity can be replicated in other studies. Furthermore.CBT Preoccupied . or whether it is just the dismissing or the preoccupied group who constitute an extreme.PPT Preoccupied . Copyright © 2010 John Wiley & Sons. the present finding that client score on the AAI unresolved scale was related to the length of speaking turns suggests that even though unresolved status is not generally related to a particular overall narrative style. whether secure clients are indeed ‘midrange’ as would be expected. these different styles might just be different ‘flavouring’ of the therapy interaction with no immediate relevance to treatment outcome. the main value of the present study lies in the guidelines it provides for further research in the area. it would be important to investigate whether clients with low and high verbal productivity generally have poorer treatment outcomes. To evaluate the possible role that verbal productivity might play in an informal process evaluation of client attachment. Furthermore. some additional issues would need to be addressed to make sense of this difference and evaluate its clinical relevance and utility. Nevertheless. In the present study. it does contribute to the distinction between types of insecurity. it would be important to know. respectively. In spite of the low statistical power. which suggests that it would be sufficient to sample a few sessions from each dyad to test this finding in a larger sample. who may mainly differ in their degree of narrative coherence.Attachment and Narrative Process in Therapy 100% 90% External mode percentage 80% 70% 60% 50% 40% 30% 20% 10% 0% 2 4 6 8 10 Session 12 14 16 18 20 Dismissing . Psychother. the client population was very select. The dependent variables analysed are likely to be affected by other client variables than attachment. It seems reasonable to interpret the two extremes with respect to verbal productivity as ‘underproductive’ and ‘overproductive’. Clin.PPT Dismissing . This should be replicated and expanded in larger and more varied client samples. since the small sample size renders the study vulnerable to strong influences from idiosyncratic patterns of the particular client-therapist dyads included. Thus. The verbal productivity difference between dismissing and preoccupied clients lends some support to the idea that attachment-related differences in discourse style are also detectable in psychotherapy sessions. whereas. Psychol. with a middle range of productivity representing the most favourable therapeutic process. which necessarily limits generalizability. verbal productivity in terms of words uttered per minute and length of speaking turns were relatively stable from session to session in each dyad. Linear prediction of percentage of external (versus reflexive) narrative process across sessions for client attachment states of mind in each of two therapy types DISCUSSION This exploratory study examined a small number of therapy dyads (N = 8) in relative detail. a number of significant differences between therapies with dismissing and preoccupied clients were detected. it may still be relevant in relation to in-session narrative process. It is quite probable that verbal productivity does not contribute much to the distinction between secure and insecure clients.

but they may be at risk of overwhelming therapists and may leave little room for reflection. concrete realities. obstructive or neutral (Frankel et al. If dismissing clients show less initiative. confused. clients with different insecure attachment states of mind may be difficult in very different ways. In the present study. I. Conversely. Psychol. The present study was inspired by the premise that although insecure clients are generally more difficult to treat than secure clients. the relevance of the NPCS to future studies on in-session narrative correlates of attachment states of mind has to be evaluated in the light of the considerable resources required to code a substantial number of therapy sessions from each dyad in a larger sample than the one used here. Thus. speak less and with shorter speaking turns. This might be related to the therapy forms. In. F. 2006). they generate less information on their experience. PPT. Slade (1999. 2000) and non-verbal emotion expression. it would be important to evaluate whether the silences exhibited by dismissing clients are reflective or perhaps disengaged. This may make it harder for therapists to attune empathically and may also generate doubts about the clients’ commitment to treatment. Further studies on attachment states of mind and pauses should address qualitative aspects of silences to make sense of the quantitative findings. The difference between dismissing and preoccupied clients in CBT might be a result of dismissing clients electing to focus more on the tangible aspects of the eating disorder. and does not easily lend itself to large-size studies. 1999). On the other hand. Psychother. it may also reflect a greater reluctance to self-disclose. whereas. The client group might also be important since eating disorders have been linked to difficulties in verbalizing feeling states (Speranza et al. the NPCS is labour intensive. The finding that a dismissing state of mind was associated with more and an increasing number of pauses across the 6 months of therapy deserves further investigation. whereas. angry. which was also very rare in the material.508 Even if the variables studied here do not affect treatment outcome. which only require transcription of sessions to be computed. since neither PPT nor CBT have the same explicit focus on the detailed description and differentiation of feeling states that is found in process–experiential psychotherapy. 588) argued that dismissing clients are likely to make therapists feel ‘caught in the same barren landscape as the patient’ and ‘utterly shut out’.g. S. preoccupied clients engage in more psychological considerations. the present study suggests that future studies inquiring into in-session emotionality related to client attachment states of mind should probably include a wider range of emotional expression. including both simple use of emotion words such as that captured by computerized dictionaries (Buchheim & Mergenthaler. secure clients should be included for comparison. Ltd. reflective and disengaged silences that may be evaluated as either productive. However. Compared with the verbal productivity variables. whereas. which distinguishes between different categories of silence such as e. which would suggest defensiveness. preoccupied clients make the therapist feel ‘swamped. Generally. Thus. the distribution of narrative process modes varied just as much from session to session within dyads as it did between dyads. Such experiences may be partly related to clients’ patterns of verbal productivity and narrative initiative. where larger proportions of internal process have been reported (Levitt & Angus. however. 498–511 (2011) . especially as treatment progresses. With regards to the relative predominance of the external and reflexive narrative process modes. The relative absence of pauses with preoccupied clients may reflect a tendency to ‘fill’ the sessions with verbal activity. it was tentatively hypothesized that dismissing clients would focus more on the external mode relating to descriptions of outer. 18. it may also reflect more emotional engagement. helpless. e. which indicates that sampling just a few sessions from each dyad would be likely to misrepresent the narrative process. If preoccupied clients show more initiative. although less pronounced. they may still affect therapist reactions to clients. where the proportion of external process tended to be larger with dismissing clients than with preoccupied clients. most of the external process did not focus on the eating disorder. 2005).. but rather on discussion Clin. they give therapists more material to work with... Furthermore. preoccupied clients would focus more on the reflexive mode relating to inner or psychological realities. leaving no room for reflection. The results of the present study suggest that it would be relevant to investigate whether therapist reactions to clients with different attachment states of mind are mediated by client in-session discourse style. The difference that had been expected for internal narrative process could not be evaluated because of low coding reliability for this mode. secure clients would be expected to show more productive and fewer obstructive silences than dismissing clients. This seemed to be the case in CBT. emotional. the higher number of pauses with dismissing clients may evidence more reflection. However. and perhaps a rise in defensiveness as therapy increasingly pushes for involvement. but the tendency in PPT was in the opposite direction. speak more and for longer periods.. verbal productivity. p. Even though the difference detected in terms of narrative initiative is theoretically interesting. This might be done using instruments such as the Pausing Inventory Categorization System. the coding of narrative process modes generated results that related more clearly to therapy type differences than to attachment states of mind. but it would be relevant to evaluCopyright © 2010 John Wiley & Sons. In her much cited chapter on attachment states of mind and psychotherapy. and dysregulated’. Daniel ate whether they are more similar to dismissing or preoccupied clients in terms of overall number of pauses.g.

& Gottman. 353–369. ed. & van Ijzendoorn. applying a detailed level of discourse analysis across a range of discourse contexts. coherence and emotional valence of the in-session autobiographical memory narratives might very well reveal differences. attachment states of mind are evaluated based on a composite of different discourse markers. Acts of meaning.. Thousand Oaks. On the other hand. 29(5). Bouffard. 18. Actual minds. Emotion.K. Perhaps. 57(5). London: Pimlico.. (1969/1997).H. M. and non-experts. The concept of coherence in attachment interviews: Comparing attachment experts. (1999). Beck Depression Inventory—Second edition Manual. Lewin. Archives of General Psychiatry. Journal of Clinical Psychology. but are expressed in a wider range of contexts (e. 87–101). Bakeman.g. Narrative Processes Coding System. The power of the narrative: A multiple code account. Bowlby. In J.J. Loss.. Canada. Angus. Main et al.. & Brown. The only variable for which a Client Attachment × Therapy Type interaction effect was noted was the relative proportion of external (versus reflexive) process in sessions. Fairburn. and the rest of the Bulimia Trial research team. M. & Levitt. Angus. possible worlds. H. Cambridge. 870–879. & Van Ijzendoorn. Pennebaker (Ed. whereas. A multicenter comparison of cognitive-behavioral therapy and interpersonal psychotherapy for bulimia nervosa. In spite of the marked difference between the relationship-focused PPT and the symptom-focused CBT. verbosity). Angus & J.). In L. As already mentioned. Attachment and loss: Vol. Walsh. Psychol. ACKNOWLEDGMENTS The author would like to thank Susanne Lunn. future studies should utilize more fine-grained discourse analysis. 1. (1986). Hardtke. R. which is based on the Gricean maxims incorporated in the AAI coding system (Grice. it requires more statistical power to detect an interaction effect than to detect main effects. but the estimated coefficients of the interaction term were also small for most variables. J. J. (1996).. Since client attachment states of mind might be expressed more clearly in therapies that focus on attachment relationships. J. The Narrative Processes Coding System: Research applications and implications for psychotherapy practice. 1255–1270. York University. it may be that engagement in psychotherapy inherently activates the attachment behavioural system. More elaborate and detailed information on the ‘linguistic fingerprints’ of attachment states of mind Copyright © 2010 John Wiley & Sons. nevertheless. A special thank you to Lynne Angus and the dedicated NPCS coders. J. B. & Rotondi-Trevisan. often involving relational experiences. London: Pimlico. (1997). investigations of the specificity.J. and research (pp. Stig Poulsen. & van Ijzendoorn. the lack of Client Attachment × Therapy Type interaction effects in this study may also be related to the particular choice of dependent variables.C. T. & Clin. perhaps including a consideration of the content of the therapeutic dialogue rather than focusing only on narrative modes. Obviously. New York: Cambridge University Press.. Observing interaction: An introduction to sequential analysis (2nd.T. 2008). Department of Psychology. & Hardtke. K.Attachment and Narrative Process in Therapy of episodes from the client’s life. J. Ltd. Inc. D..1. ‘What’s the story?’ Working with narrative in experiential psychotherapy. C. Unpublished manuscript. (1996).. (2000). The handbook of narrative and psychotherapy: Practice. London: Pimlico. Levitt. (1993). Psychother. Separation. G. H.. 459–466.A. General Sequential Querier for Windows (Version 4. Bruner. (2006).D. (1990). J. MA: Harvard University Press. K. (1980/1998). & Quera. 2006). 1975. The results of the present study suggest that the different insecure attachment states of mind may not be differentiated by distinct in-session narrative process at the relatively broad level picked of by the coding for NPCS narrative process modes.. San Antonio.. This kind of process seemed equally predominant with dismissing and preoccupied clients. M. Attachment and loss: Vol. Developmental Psychology.. Beijersbergen. linguists. theory.T. H.. Bowlby.). (1995). Bakermans Kranenburg. REFERENCES Agras. 509 in different contexts would contribute importantly to our understanding of attachment states of mind and their psychotherapeutic implications. Attachment. Training Manual. Bruner. W. the difference between client attachment states of mind was larger in CBT than in PPT. Bakermans Kranenburg.. preoccupied anger). some of the markers relied upon in AAI classification depend more narrowly on activation of the attachment behavioural system (e. The topic is worthy of further investigation. 498–511 (2011) .. and in this case. 8(4). 3. North York. Attachment and Human Development... Bakeman. Cambridge. Wilson. Steer. In the AAI. disclosure. Bowlby. Attachment and loss: Vol. Beck. other characteristics may be more general discourse traits that are also related to specific attachment histories.M. 2.5). L. A psychometric study of the Adult Attachment Interview: Reliability and discriminant validity. Bucci. W. L. Ontario. J. including the AAI proper. neutral conversations and different kinds of psychotherapy.g. Thus.. M. MA: Harvard University Press.. V. A.. Such investigations might utilize an instrument such as the recently developed Coherence Q-sort (Beijersbergen. R. CA: Sage Publications. G.). Bakermans Kranenburg. most of the differences between dismissing and preoccupied clients identified in this study were similar in both therapies. even when the therapy dialogue does not directly address attachment.E.G. (1973/1998). M. (2004). L. Angus.H.. (2007). W. sessions from two relatively different therapy forms were investigated.S. 55(10). McLeod (Eds. & Kraemer. R. TX: The Psychological Corporation.

R. Psychother. Logic and conversation. Treboux. Unpublished manuscript. In H. Speech acts (pp. Jaffe. N. New York: Guilford Press. (2003). 49–66). O. New York: Guilford Press. H. G.L. I. MN: National Computer Systems. 615–625. & Lee. F. 3(1). C.. Psychotherapeutic issues on eating disorders: Models. M. Steele. and clinical applications (pp. London: Brunner Routledge. 16(5). S. Kennedy. Goldberg & R.F. The relevance of attachment research for adult narratives told in psychotherapy. (2008). Journal of Nonverbal Behavior. 47–60.. Grice. 395–433)..C. J.. Dozier.. In J. Dozier. 69–98). Clinical Psychology Review. 390– 407. Cassidy & P. Cartwright. K. (1994). (1998). Fonagy. & Main.L. Handbook of attachment: Theory. Leigh.A. London: Lawrence Erlbaum Associates.. Fonagy. H. Minneapolis. M. Lunn. Steele. (1999). 74(6). M. B. L.H. & Kernberg.R. Monographs of the Society for Research in Child Development. The relation of attachment status. Psykoanalytisk psykoterapi med bulimi—En behandlingsmanual. Jurist.. Cole & J.. C. S.. Ltd. Ciocca (Eds. Code-A-Text Integrated System for the Analysis of Interviews and Dialogues (Version 3.D. Applied Developmental Science. H. G.. Target. mentalization. Z. L.. Psychotherapy process measure research and the evaluation of psychotherapy orientation: A narrative analysis. M. Attachment and psychopathology in adulthood. England: Analytic Press.. Levitt.S. E. Practice. York University.)... (1996). J. Cognitive behavior therapy and eating disorders.. D. Morgan (Eds. Rome: Societa Editrice Universo.. research. vi–131. Murray. Journal of Psychotherapy Integration. van Ijzendoorn. 139–153). K. M. 233–278). Ellgring. The developmental roots of narrative expression in therapy: Contributions from attachment theory and research. In G. Atkinson & S. Canada. Adam.. 46(3). Korfmacher. E. E. J. 62(2). [Psychoanalytic psychotherapy for bulimia—A treatment manual]. Fairburn. S. Mattoon.). 43–52. 93–122).). Steele. Child Development. (2008). The relationship among attachment representation. Cognitive behaviour therapy versus psychoanalytic psychotherapy for bulimia nervosa—A randomised. Research.L. 9(3). Copyright © 2010 John Wiley & Sons. Kaplan. Hesse. Kennedy. Shaver (Eds. Crown.. T. H. In J.B. Holmes. Syntax and semantics. & Belsky... 373–384. ColonDowns.E.M. 66(2).. New York: Guilford Press. & Jasnow. Meehan.R. P. & Target. Psychol. (2006).. Rhythms of dialogue in infancy: Coordinated timing in development. Steele (Eds. A. Mattoon.R. 1027–1040. and the development of the self. 279–300.N. & Angus. Hillsdale. Adult Attachment Interview (3rd ed. Waters. Unpublished thesis. and clinical applications (pp. (1996). Tyrell. L. scoring and procedures manual (3rd ed. R. 20(2). Attachment organization and treatment use for adults with serious psychopathological disorders. K. Reynoso. (1991). M. Steele & M. Clinical applications of the Adult Attachment Interview (pp. Washington DC: American Psychological Association. S. J. Attachment and Human Development. Shaver (Eds.). The challenge of treatment for clients with dismissing states of mind. L. & Feider. (2001). Buchheim. M. Healing stories: Narrative in psychiatry and psychotherapy (pp. 498–511 (2011) . NJ. E. 167–180). 10(4). SCL-90-R: Administration. Change in attachment patterns and reflective function in a randomized control trial of transference-focused psychotherapy for borderline personality disorder.S.. Steele & M.).M. University of California at Berkeley. P. 1(1). A. Dozier.. Levy.C. Fish. University of Copenhagen. Dozier. 497–519).M. & Bakermans Kranenburg..). E. (2001). (2006). Department of Psychology. Derogatis. London: Oxford University Press... & Bates. research. Ogawa. Adult attachment: Implications for the therapeutic process in a home visitation intervention.. K. J. M. In H. Weber. Gergely. 62–76. Vocal indicators of mood change in depression. (2005). & Poulsen. (1995). A comparative analysis of the Narrative Process Coding System and three standardized psychotherapy process measures: A multi-modal analysis. B.. S.6. L. C. Lomax. Greenberg. L. & Egeland. Assessing silent processes in psychotherapy: An empirically derived categorization system and sampling strategy. R. K. developmental. In S. Daniel.I. H. The adult attachment interview: Historical and current perspectives. methods. Candelori.. Daniel Frankel. J... Defensive and creative uses of narrative in psychotherapy: An attachment perspective. K.. and borderline states: The predictive specificity of the Adult Attachment Interview and pathological emotional development. (1993). M. 27(1). (2004). Attachment. J.. New York: Guilford Press.. (1996). J.510 health (pp. New York: Other Press. D. The Bulimia pilot trial. Handbook of attachment: Theory. Volume 3. Stovall. (1990).C. M. (1999). Bria & A. (2000). New York: Guilford Press. Feldstein.. Journal of Consulting and Clinical Psychology. K. M. 41–58). and response to psychotherapy. Crowell. 26(8). and narrative style: A computer-based text analysis of the Adult Attachment Interview.J. S. 301–316. emotion-abstraction patterns. (1975). C.. psychiatric classification. Clarkin. Department of Psychology. C. Attachment state of mind and the treatment relationship. & Albus. & Scherer. & Poulsen. Clin. Adult attachment patterns and individual psychotherapy: A review. A. P. Levitt. Psychotherapy Research. Unpublished manuscript. Journal of Consulting & Clinical Psychology. 64(1).). Development and Psychopathology. (2002). & Gerber. Interactional synchrony and the origins of infant-mother attachment: A replication study. Holmes (Eds. G. Affect regulation. (1999).. T. & Levine. 968–984. 18.. 99–125). & Angus.. 83–110. R.). Steele. B. Holmes.E. Attachment theory and psychotherapy. In P. S. S. B. (2006). The search for the secure base.. Attachment and eating disorders. the reflective self...).. Isabella. Attachment issues in psychopathology and intervention (pp. Unpublished manuscript.. (2004).A. Daniel. 27–40. Steele (Eds. O’Connor. Beebe.. & Dudas. observerblinded clinical trial. J. Muir (Eds. Leigh.P. Kelly. M.. Lunn. George.. Publishers. Discriminant validity of the Adult Attachment Interview.0).M. 22–31. 2(1).). Psychotherapy Research.. The distribution of adult attachment representations in clinical groups: A meta-analytic search for patterns of attachment in 105 AAI studies. In L. (2001).). Goldberg (Eds. O. Attachment theory: Social. (2008). Heinicke. M.S.F.I.... H. The AAI anticipates the outcome of a relation-based early intervention. Cassidy & P.. & Target. 2584–2599.F. (1997). 67(5). Child Development. C. New York: Academic Press.. Training. (1999). & Mergenthaler. M. Levitt. In P. Psychotherapy: Theory. Clinical Applications of the Adult Attachment Interview (pp. A. Fonagy.. M. S. and clinical perspectives (pp. (1999). Inc. C. H. and results (pp. (1996). & Ciocca.W. (2009). Roberts & J. Clinical Social Work Journal.F. E..

. B.. J. D. Steele. Goldberg & R.. Studying differences in language usage in recounting attachment history: An introduction to the AAI... A. Venisse.).A. The development of emotion expression during the first 2 years of life..). Magai.. M. E. Mikulincer... (1999).. Linking verbal and nonverbal representations: Computer analysis of referential activity. University of California at Berkeley.. Parish. Goldwyn. & Janssens. research. Handbook of attachment: Theory.). J. 2(2). Schuengel. In H.. Teti. British Journal of Medical Psychology. M. C. M.1. Department of Psychology. New York: Guilford Press. New York: Guilford Press. Main. J. P. Tyrrell. 375–388. Adult attachment. (Eds. Structured clinical interview for DSM-III-R personality disorders (SCID-II). D. and clinical applications (pp. Rabe-Hesketh.H.. Stability and discriminant validity of the Adult Attachment Interview: A psychometric study in young Israeli adults. R. F. Sorbello. S. J. Teague. Psychol. Hillsdale. J.. (2009).P. with selected implications for clinical work. Patrick. New York: Guilford Press. & O’Connell. 497–505.L.M. (2002). Steele (Eds. A. (2007). Steele & M. (1994). Killeen. Attachment patterns among young women diagnosed with bulimia nervosa and their mothers. 1055–1096. 6(2). Clin. D.H. (1994). inflexible attention under attachment-related stress. Psychotherapy relationships that work: Therapist contributions and responsiveness to patients. (1990). TX: Stata Press. E. S. Dozier.. & Bucci. C.H. Speranza. Multilevel and longitudinal modeling using Stata (2nd ed. 407–474). 1–104. Pazzagli.. New York: The Guilford Press. 67(5). 48(4).. M.. Sagi. M. Clinical Applications of the Adult Attachment Interview (pp. Main. Personality disorder and the mental representation of early social experience. R.. P. Sandros.N. C. & Shaver.. 787–802).).. & Eagle. Hesse. Psychotherapy Research. O. N. In H. Flament. van IJzendoorn.. M. Ltd.A.J. & Pallanti. (1989).A. & Stinson. R D. Main. (2007). Malatesta. C. (2003). 339–354. dynamics.. DC: American Psychiatric Press. Gibbon. C. Perez Diaz.A... imagery. Muir (Eds. A.. Hess. (2008). University of Copenhagen.. 498–511 (2011) .. 31–68). Handbook of attachment: Theory. developmental. (2002). F. R. & Maughan. Psychiatry Research. Attachment representations of institutionalized dolescents and their professional caregivers: Predicting the development of therapeutic relationships. D. & Hesse. & Cicchetti. L. Candelaria. In J.L. Effective treatment relationships for persons with serious psychiatric disorders: The importance of attachment states of mind. Psychotherapy transcription standards. (1992). In J. New York: Guilford Press. C. Loas. (1999)... Howard. 76(3).Attachment and Narrative Process in Therapy University Clinic. Castle. Washington. Ramsay. The organized categories of infant. 54(1–2).. Mergenthaler. (2008). and clinical applications (pp. Unpublished PhD thesis. In H.. British Journal of Medical Psychology. M. & Shepard. M. (2001).. Williams. Inc. R. L. M. M. & Berant. Attachmenttheory-informed intervention and reflective functioning in depressed mothers. Stephan. Culver. A.). Ward. J..M. 18.. College Station. W. American Journal of Orthopsychiatry..L. In S. M. and attachment: Working models of interpersonal affect and the socialization of emotion. 154–174). (1999). R.. C. & Skrondal. R.). Cassidy & P. Zegers. 135(2). Slade.. M.W. Department of Psychology. (2008).. 325–334 Copyright © 2010 John Wiley & Sons. Attachment processes in eating disorders. Scharf.. Psychoanalytic Psychology. Eating and Weight Disorders. O.. M. E.Z. (1999). S. Halfon. B. (2006).B. Turnbull. Unpublished manuscript.H. (2005). Toth. Steele (Eds. Developmental Psychology. Rogosch.. 511 Ramacciotti. and change. 575–594). New York: Guilford Press. University of Copenhagen. Journal of Consulting and Clinical Psychology. and clinical perspectives (pp. 74(4). Wallin.. W. (2001). M. (2008). Shaver (Eds. Norcross. 72(3)..R.. 166–170. Steele & M. 153–163. Attachment theory and research in clinical work with adults. Steele (Eds. J.. Cassidy & P. Vismara. and developmental outcomes in an African American sample. (2006). Department of Psychology. Tesman. van Ijzendoorn. A. Monographs of the Society for Research in Child Development. Affect. Bizouard. 126– 153). R...B.C. and adult attachment: Flexible vs. parental commitment to early intervention. Corcos.M. & Goldwyn. NJ: Analytic Press. Depressive personality dimensions and alexithymia in eating disorders. 125–142. Clinical Applications of the Adult Attachment Interview (pp. G. M. Adult Attachment Scoring and Classification Systems. Obegi. & Treasure. Psychother.. R. Attachment in adulthood: Structure. research... Recent studies in attachment: Overview. S. P. A. 771–777. Attachment to the therapist.. Attachment in Psychotherapy.. child. M.M. New York: Guilford Press. Hobson. E. C.). 725–733.R. (2000). 30(5). Attachment theory and research: Implications for the theory and practice of individual psychotherapy with adults. & Koren-Karie. M. M. Attachment theory: Social. Spitzer. A. Version 7. Attachment in anorexia nervosa: A transgenerational perspective. G.L. M. H. Journal of the American Psychoanalytic Association. & Jeammet. Development and Psychopathology. Shaver (Eds. M. Guilbaud. New York: Guilford Press. 271–286. 20(2). Miller. Steele & M.R. Main. M.. New York: Oxford University Press. E.. 6. M. Steele. (1995).B. Mergenthaler.. P. Mancone. Clinical applications of the Adult Attachment Interview (pp. & First.). & Fallot.

Copyright of Clinical Psychology & Psychotherapy is the property of John Wiley & Sons. download. . Inc. users may print. However. and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. or email articles for individual use.

Sign up to vote on this title
UsefulNot useful