Psychotherapy, Vol 63(3), Sep 2026, 284-294; doi:10.1037/pst0000623
Research on microprocesses underlying alliance formation remains limited compared to outcome-based studies, particularly in videoconferencing psychotherapy (VCP). This study examined therapist and client interaction patterns predicting early emotional bond in VCP and face-to-face therapy. A total of 56 early therapy sessions from 28 therapist–client dyads were coded with the Therapeutic Relationship Coding System, yielding 35,715 behaviors. Alliance strength was determined through an extreme-groups design based on the Bond subscale of the Working Alliance Inventory–Patient version. Bayesian logistic regression models assessed whether therapist behaviors and therapist–client sequences predicted alliance and their interaction with modality (VCP and face-to-face). Empathy alone predicted weaker bond in VCP, whereas sequences where it was followed by clients’ agreement predicted bond more strongly in VCP. Active listening followed by client agreement also predicted stronger bond only in VCP, while both behaviors followed by client disagreement predicted weaker bond across modalities. Joking and accurate recall of prior client information were particularly associated with strong bond in VCP, whereas early therapist disagreement predicted weaker bond across modalities. Findings indicate that precise, accurate, and context-sensitive verbal interactions are especially critical for fostering emotional bond in VCP, where nonverbal cues are limited. (PsycInfo Database Record (c) 2026 APA, all rights reserved)