Affective and Goal-Oriented Factors of Relationship Formation in the Digital Therapeutic Alliance: A Longitudinal Study of Mental Health Chatbots

Key findings from this study

This research indicates that:

• User-chatbot relationship formation is driven primarily by emotional support and practical assistance rather than trust or satisfaction as independent drivers.
• Conversational control functions as a measurable predictor of relationship quality alongside affective and goal-oriented factors.
• Trust and satisfaction operate as outcomes of supportive, effective interactions rather than standalone antecedents of relationship formation.

Overview

This longitudinal study examined relational mechanisms underlying mental health chatbot effectiveness by operationalizing a multi-dimensional instrument to assess user perceptions of digital therapeutic alliance. The research involved 56 participants interacting with two widely-used CBT-based chatbots (Wysa and Youper) over four weeks.

Methods and approach

Researchers conducted a within-subjects design with iterative factor refinement and regression modeling. Participants engaged with mental health chatbots while researchers captured perceptions of relational and functional dynamics through structured instrumentation. Analysis examined which interpersonal and functional variables predicted user-chatbot relationship formation.

Results

Two primary factors drove user-chatbot relationship formation: an affective factor centered on emotional support and a goal-oriented factor centered on practical assistance. Conversational control contributed as an additional predictor alongside these interpersonal dimensions. Trust dimensions (privacy and non-judgmentalness) and satisfaction emerged as correlated outcomes of supportive, effective interactions rather than independent predictors of relationship quality. These findings clarify the underlying structure of digital therapeutic alliance in conversational agents.

Implications

The dual-factor model of relationship formation has direct design implications for mental health chatbots. Developers should prioritize balancing empathy and efficacy rather than treating these dimensions as separable design concerns. The finding that trust and satisfaction function as outcomes rather than drivers suggests interventions should focus on strengthening affective and goal-oriented interactions to indirectly enhance perceived trustworthiness and user satisfaction.

These results advance theoretical models of how users form relationships with conversational agents in clinical contexts. The identification of specific interpersonal and functional mechanisms enables more precise measurement and iterative refinement of digital therapeutic alliance instruments. Future work should examine whether the identified factor structure generalizes across different chatbot architectures and clinical populations.

Scope and limitations

This summary is based on the study abstract and available metadata. It does not include a full analysis of the complete paper, supplementary materials, or underlying datasets unless explicitly stated. Findings should be interpreted in the context of the original publication.