Ambivalence and Alliance Ruptures in Cognitive Behavioral Therapy for Generalized Anxiety

Abstract

Client ambivalence about change (or motivation) is regarded as central to outcomes in cognitive behavioral therapy (CBT). However, little research has been conducted to examine the impact of client ambivalence about change on therapy process variables such as the therapeutic alliance. Given the demonstrated limitations of self-report measures of key constructs such as ambivalence and motivation, the present study instead employed a newly adapted observational measure of client ambivalence. Client statements regarding change (change talk (CT) and counter-change talk (CCT)) were coded in early (session 1 or 2) therapy sessions of CBT for generalized anxiety disorder. The frequency of CT and CCT was then compared between clients who later experienced an alliance rupture with their therapist, and clients who did not. The results showed that clients in dyads who later experienced an alliance rupture expressed significantly more CCT at the outset of therapy than clients who did not later experience an alliance rupture. However, CT utterances did not significantly differ between alliance rupture and no-rupture groups. CCT may strain the alliance because clients expressing higher levels of CCT early in therapy may be less receptive to therapist direction in CBT. Consequently, it is recommended that clients and therapists work together to carefully address these key moments in therapy so as to prevent alliance rupture and preserve client engagement in therapy.

Keywords::

• ambivalence
• alliance rupture
• observational measure
• change talk
• counter-change talk

Notes

^1. We have chosen this measure as opposed to the usual method for quantifying client motivational talk (CT/(CT+CCT)) for a number of reasons. First, due to the CBT context, the therapists were not deliberately eliciting motivational language as they would have in the MI setting in which this formula was developed. Thus, the overall proportion of CT and CCT in the context of all utterances is relevant. Second, importantly, our measure controls for overall client verbosity, whereas the original index does not. Last, our measure allows for CT and CCT to be measured independently of one another (rather than simply being the inverse of one another), consistent with our study aims.

^2. Note that the results are the same when using only those 29 subjects with the session 1 videotape available, i.e., CCT: t(27) = 2.09, p = .044; CT: t(27) = 0.52, p = .581.