Perceptions of psychotherapeutic response modes: The role of attachment style

This analog study explored a theoretical link between attachment style and reactions to psychotherapy interventions. The level of intervention chosen for analysis was the verbal response mode. Excerpts were edited and spliced from full-length demonstrations of different types of psychotherapy in such a way that various response modes were grouped in theoretically meaningful ways: responses clarifying the therapist's expectations, those more ambiguous about the therapist's expectations, those providing general supportiveness, and those conveying overt approval. Subjects' scores were obtained after they viewed each of six excerpts--two contrasting excerpts for each of three therapist/patient dyads. Specific scoring differences were predicted for four subscales from the Checklist of Psychotherapy Transactions-Revised (CLOPT-R: Kiesler, Goldston & Schimdt, 1991): dominance, submission, friendliness, and unassuredness. The predictions were based on the currently known impact of specific response modes and on the characteristics of individuals falling into each of four currently accepted categories of adult attachment style. For example, subjects classified as Fearful and Preoccupied, because of their high need for approval, were predicted to rate responses that clarified therapist expectations as more assured and friendlier than responses that left therapist expectations ambiguous compared to Secure and Dismissing subjects. There were significant differences in the way subjects scored the response modes (main effect for response mode) for only two of the three contrasts. Response mode and attachment style interacted (interaction effect) significantly for the targeted dependent variables only for the contrast between generally supportive response modes and overtly approving ones. As predicted, there was a group difference along the submission/dominance axis; however, it was not in the predicted direction. Moreover, when general neuroticism was included as a covariate, the interaction lost significance. Secondary analyses revealed unpredicted group differences but no clear pattern. Methodological and theoretical reasons for the unexpected results are discussed, including the possibility that attachment schema may not have been activated in subjects or that attachment may be relationship specific rather than a global personality trait.