Distinct roles of state-like and trait-like patient–therapist alliance in psychotherapy

Sigal Zilcha-Mano, Hadar Fisher

Research output: Contribution to journalReview articlepeer-review


Hundreds of studies suggest that the patient–therapist alliance is the most consistent predictor of treatment outcome across patients, therapists, types of treatment and settings. Yet, the context-insensitive nature of this prediction makes it difficult to optimize alliance to benefit individual patients. To elucidate the potential roles of alliance in treatment, research published mainly in the past five years has distinguished between state-like strengthening of alliance throughout treatment from trait-like differences between individuals in alliance strength. In this Review, we synthesize the literature on state-like and trait-like alliance to shed light on for whom, when and why alliance predicts treatment outcome. Trait-like alliance is a product of patients’ and therapists’ intrapersonal and interpersonal characteristics, whereas state-like alliance reflects the therapeutic process evolving between them. Trait-like alliance mediates the effects of patient and therapist baseline characteristics on outcome, whereas state-like alliance is the mechanism by which alliance might drive therapeutic change. This synthesis reveals the importance of putting alliance into context for making progress toward personalized psychotherapy, in which the alliance is not merely a non-specific factor but an individual-specific mechanism of change.

Original languageEnglish
Pages (from-to)194-210
Number of pages17
JournalNature Reviews Psychology
Issue number4
StatePublished - Apr 2022

Bibliographical note

Publisher Copyright:
© 2022, Springer Nature America, Inc.

ASJC Scopus subject areas

  • Psychology (miscellaneous)
  • Clinical Psychology
  • Developmental and Educational Psychology


Dive into the research topics of 'Distinct roles of state-like and trait-like patient–therapist alliance in psychotherapy'. Together they form a unique fingerprint.

Cite this