Grounding of the AI-Based Interface

What it measures

This scale measures how well a person believes their communication was received and understood by a particular AI-based interface. Bergner, Hildebrand, and Häubl (2023) described the measure as adapted from Dillenbourg and Traum (2006), but that article contained no scales; it is therefore more precise to describe the measure as created by the authors, drawing on grounding concepts from that earlier work.

Response format

7-point Likert (1=Not at all, 7=Fully agree)

Scale items

• The system provided feedback of having perceived my input.
• The system provided feedback of having understood my input.
• The system provided feedback of having accepted my input.
• I felt this system understood what I had to say.
• I feel like the system acknowledged and confirmed what I had to say.

Full scale items shown. Always cite the original source when using this scale in research.

Usage notes & scoring

Alphas were .89 and .91 in Studies 3 and 4, respectively (Bergner, Hildebrand, and Häubl 2023, pp. 755, 758). Validity was not discussed per se, but the scale was used as a successful manipulation check in both studies, offering preliminary support for predictive validity. Not a true item-level adaptation of Dillenbourg and Traum (2006) — see About section.

Citation

Bergner, Anouk S., Christian Hildebrand, and Gerald Häubl (2023), "Machine Talk: How Verbal Embodiment in Conversational AI Shapes Consumer-Brand Relationships," Journal of Consumer Research, 50 (4), 742-764.

Keywords