Attachment Theory and Psychotherapy Research

John Bowlby's^1,2 attachment theory has had a profound influence on developmental psychology, but until the past decade it has had a limited influence on psychotherapy research.³ Recent years have seen a great increase in psychotherapy research that draws explicitly from attachment theory. For example, studies of persons with and without serious psychiatric disorders show that attachment states of mind are associated with different approaches to interpersonal relationships.⁴ Furthermore, individual differences in clinicians' attachment states of mind appear to influence the treatment relationships they form.⁵ This increased interest among psychotherapy researchers is reflected in two journals' recent special issues dedicated to attachment^3,6 and also in the several chapters devoted to psychotherapy in the recently published Handbook of Attachment.⁷

In part, the recent interest is the product of methodological improvements in measuring and classifying attachment styles.^8–12 One classification system that seeks an integration of other models is that of Bartholomew and Horowitz.^8,13 They posit two orthogonal dimensions that in combination yield four main attachment styles. One dimension measures self-image along a negative or positive continuum; the other measures one's positive or negative image of others. Individuals with a “secure” attachment (positive self, positive other) are presumed to have internalized a sense of self-worth and a trust that others will be available and supportive; they are autonomous, yet comfortable seeking and expecting support from others. “Preoccupied” individuals (negative self, positive other) are preoccupied with attachment needs and depend overly on others for personal validation, acceptance, and approval. “Fearful” individuals (negative self, negative other) view others as uncaring and unavailable; they view themselves as unlovable. “Dismissing” individuals (positive self, negative other) distance themselves from others, viewing themselves as self-reliant and invulnerable to rejection by others.

In this set of research abstracts, I present a selection of the psychotherapy research on attachment that has been published in the last two years. First, a brief summary of attachment theory is in order.

John Bowlby posited attachment-seeking as a primary human motivation, rejecting the orthodox psychoanalytic view that attachment is the derivative of sexual or oral instinctual drives. The essence of Bowlby's attachment theory is the proposition that affectional bonds between individuals and patterns of early life interactions between caregivers and children produce internal working models that serve as templates guiding interpersonal expectations and behaviors in later relationships. Caregivers who are stable, consistent, and predictable tend to encourage the development of internal working models of the self as valued and others as trustworthy and reliable sources of nurturance. Unstable, inconsistent, or unpredictable caregiving in early life can produce maladaptive internal working models that are reflected in insecurity and anxious forms of attachment.

ABSTRACTS

Tyrrell CL, Dozier M, Teague GB, Fallot RD: Effective treatment relationships for persons with serious psychiatric disorders: the importance of attachment states of mind. Journal of Consulting and Clinical Psychology 1999; 67(5):725–733

Summary: The authors examined whether treatment relationships and client outcomes are affected by the interpersonal attachment styles of clinical case managers and their clients with severe psychiatric disorders. The authors suggest that differences in treatment relationships mirror differences in how individuals organize relationship information internally, in different states of mind, as a function of earlier attachment experiences. They hypothesize two basic attachment states of mind: deactivating and hyperactivating. Deactivating states of mind are associated with deflecting information about attachment topics in order to avoid early attachment relationship themes. They have been associated with maintaining interpersonal distance from others. Hyperactivating attachment states of mind are associated with preoccupation with attachment or relationships. Adults with these states of mind tend to admit more personal distress as compared with others and show evidence of unresolved conflict with their parents. Tyrrell and her co-authors hypothesized that case managers (n=21) whose attachment states of mind differ from those of their clients (n=54) would balance their clients' usual strategies and thereby help them better manage their emotional distress and maintain symptomatic and functional stability. They assumed that dissimilar matches would challenge clients' characteristic ways of processing emotion in relationships, thus facilitating the learning of new ways to regulate distress and manage relationships. Results supported these hypotheses. Clients who were more deactivating had better therapeutic alliances and functioned better when they had less deactivating case managers. Clients who were less deactivating worked better with more deactivating case managers, as rated by clients.

Comment: These results suggest that dissimilarity matching of clients and case managers with regard to attachment style can predict both the quality of their relationship and treatment outcome. Perhaps the specific mechanism involved is that clinicians disconfirmed their clients' usual interpersonal and emotional strategies and expectations, increasing the repertoire of clients' interpersonal behavior patterns and thus leading to greater client satisfaction.

Hardy GE, Aldridge J, Davidson C, Rowe C, Reilly S, Shapiro DA: Therapist responsiveness to client attachment styles and issues observed in client-identified significant events in psychodynamic-interpersonal psychotherapy. Psychotherapy Research 1999; 9(1):36–53

Summary: The authors content-analyzed client-identified significant events (n=10) in psychotherapy in order to understand how therapists might differentially respond to clients (n=16) with different attachment styles. By close analysis of psychotherapy transcripts, the authors found that therapists of clients with a dismissing attachment style struggled to find a coherent joint narrative that would aid in understanding the clients' early and painful attachment problems. A dismissing style is characterized by an avoidance of intimacy and conflict, a tendency to adopt the attachment figure's view in relationships, and proneness to shame and an overdeveloped sense of responsibility. The authors also identified individuals with a preoccupied attachment style, which is characterized by a high level of concern and attention to closeness and separation. These individuals exaggerate danger and because of their anxiety are unable to explore their environment productively. Therapists in the study tended to respond to individuals who had a preoccupied attachment style by avoiding interpretation, instead seeking to contain the client's emotion and to reflect upon the client's experiences. In these dialogues the therapist appeared to make a particular effort to meet the client's need to be understood, rather than challenging the client. Three attachment issues were identified in the content analyses. These were loss and abandonment, negative interpersonal feelings, and a need to be close and to feel cared for.

Comment: The content-analytic approach used by the authors does justice to the richness of the database. The findings support the authors' hypothesis that different attachment styles elicit different forms of responsiveness from therapists.

Kilmann PR, Laughlin JE, Carranza LV, Downer JT, Major S, Parnell MM: Effects of an attachment-focused group preventive intervention on insecure women. Group Dynamics: Theory, Research, and Practice 1999; 3(2):138–147

Summary: These authors explored a manualized, attachment-focused group intervention that was developed to address personal and social concerns that place insecure men and women at risk for relationship distress. The intervention sequentially addresses the following issues: atypical dysfunctional relationship beliefs; childhood attachment issues that affect later partner choices and relationship patterns; skills training in relationships; and relationship strategies. This was a highly structured psychoeducational intervention with specific group exercises. The group participants were 23 insecurely attached, never married women, most of whom were categorized on the Bartholomew system as having either a dismissive or a fearful attachment pattern. Pre- to post-intervention comparisons showed that individuals in the attachment-focused intervention, as compared with a matched control group, reported themselves as less domineering and cold and as more assertive. At a post-intervention follow-up, these differences grew stronger. In addition, those in the attachment-focused intervention rated themselves as less fearful and more secure.

Comment: These findings are based on a small sample and on paper-and-pencil self-ratings rather than behavioral observation of actual relationships. Nevertheless, they suggest that a programmatic preventive measure administered in a group format may have positive effects on later relationships. I hope the authors can find a way to test whether their intervention affects actual relationship patterns.

Kivlighan DM, Patton MJ, Foote D: Moderating effects of client attachment on the counselor experience–working alliance relationship. Journal of Counseling Psychology 1998; 45(3):274–278

Summary: This study explored the role that client attachment level plays in the relationship between a counselor's level of experience and the therapeutic alliance. The authors hypothesized that a linear relationship exists between counselor experience and the alliance for patients with more attachment problems, whereas counseling experience should be unrelated to the alliance for patients with fewer attachment problems. They used a dimensional measure of attachment that measures the degree of a client's needs for intimacy, trust in others, and freedom from fear of abandonment. They studied 40 client–counselor dyads in a university counseling center. About half of the therapists were graduate students and the others were either senior staff or predoctoral interns. A hierarchical multiple regression analysis confirmed the hypothesis. Specifically, when clients were uncomfortable with intimacy, counselor experience was positively related to the strength of the working alliance. When clients had moderate to high levels of comfort with intimacy, counselor experience did not correlate with the level of client-rated working alliance.

Comment: Studies of client characteristics that influence the alliance are useful in helping us to better understand the alliance and how therapists' level of experience influence it. The study is correlational in nature, and thus no causal inferences can be made about how client attachment factors and counselor experience affect the working alliance.

Rubino G, Barker C, Roth T, Fearon P: Therapist empathy and depth of interpretation in response to potential alliance ruptures: the role of therapist and patient attachment styles. Psychotherapy Research 2000; 10(4):408–420

Summary: These researchers explored how therapist attachment style affects the way therapists handle ruptures in a therapeutic alliance. “Ruptures” are negative shifts in the quality of the alliance, similar to “empathic failure” as conceived by Kohut. Ruptures may provide an opportunity for the therapist and patient to explore the patient's expectations in relationships and to resolve maladaptive relationship problems. In this analog study, 89 clinical psychology graduate students observed four videotaped vignettes that depicted patients displaying four different attachment styles (preoccupied, dismissing, fearful, and secure). Therapist responses were rated for level of empathy and depth of interpretation. The authors found that more anxious therapists tended to respond less empathically than those who were less anxious, suggesting an influence of therapist attachment style. As a group, the therapists tended to respond more empathically to the fearful than to the dismissing and secure patients; they were also more empathic toward the preoccupied than toward the dismissing patient. As for depth ratings, therapists tended to give deeper responses to the fearful patient than to the secure or dismissing patients. No therapist main effects for depth of interpretation were observed.

Comment: This study provides further support for the notion that the attachment styles of patients tend to elicit specific response styles from therapists. It provides marginal support for the idea that therapists' attachment styles may affect how they respond to patients.

Eames V, Roth A: Patient attachment orientation and the early working alliance: a study of patient and therapist reports of alliance quality and ruptures. Psychotherapy Research 2000; 10(4):421–434

Summary: These authors asked whether patients' attachment style is associated with the quality and development of the alliance in the early phase of psychotherapy. They also explored the relationship between attachment style and the frequency of ruptures in therapy. They studied 11 experienced therapists who provided either psychodynamic psychotherapy, cognitive-behavioral psychotherapy, or a mix of the two to 30 individuals. Data presented are restricted to the first five therapy sessions. They used the Bartholomew and Horowitz categorical model to measure patient attachment style, but in a dimensional fashion. Results showed that a more fearful attachment orientation was related to lower alliance ratings and that a more secure attachment orientation predicted higher alliance ratings. A preoccupied style was negatively correlated with the alliance, although not at a statistically significant level. The authors were surprised to find that the dismissing dimension was positively correlated with the alliance. Similar results were obtained from both client and therapist ratings of the alliance. With regard to ruptures, therapists were much more likely than patients to report them. The preoccupied attachment style predicted more therapist reports of ruptures than the other styles. A dismissing style was associated with fewer reports of ruptures both by therapists and patients.

Comment: These findings dovetail nicely with those of Kivlighan and colleagues as reported above. It would have been interesting to add therapist experience as a moderating variable in the current study. The correlations between rupture frequency and attachment style may be the first reported in the psychotherapy literature. I concur with the authors' suggestion that the high preoccupation/low dismissing relationship may be due to a higher level of tension in therapeutic relationships involving patients with preoccupying attachment styles.

CONCLUSIONS

An assumption underlying this body of research is that therapists become attachment figures for patients and that the therapeutic relationship reactivates patients' working models of attachment with regard to self-valuation and expectations of the availability and nurturance of others.² None of these studies explicitly explores the question of links between early attachment experiences of therapists and how those experiences affect the alliance. Although establishing this link seems crucial, interesting questions have nevertheless been addressed, including how clients' and therapists' interpersonal attachment styles affect therapy processes. These studies suggest that attachment theory provides a rich base for understanding the processes and outcomes of psychotherapy, particularly as they unfold through the therapeutic alliance.