Development and validation of a multidimensional scale for AI tutoring acceptance in higher education

Area of Science:

• Educational Technology
• Artificial Intelligence
• Psychometrics

Background:

• Artificial intelligence (AI) integration in higher education is growing.
• Existing tools lack comprehensive validation for assessing AI tutoring systems' perceived acceptance quality.
• Current evaluation models often overlook crucial aspects like motivation, relational experience, and trust.

Purpose of the Study:

• To develop and validate a learner-centered scale for measuring AI tutoring acceptance quality.
• To address the gap in psychometrically validated instruments for AI in education.
• To provide a multidimensional assessment of learners' experiences with AI tutors.

Main Methods:

• Literature review, expert consultation, and pre-testing were used for scale construction.
• The scale was designed with four dimensions: Feedback Quality, Instructional Effectiveness, Interaction Experience, and User Trust.
• Data were collected from undergraduate students across five universities in China.

Main Results:

• Exploratory Structural Equation Modeling (ESEM) and bifactor modeling confirmed a robust bifactor structure.
• Reliability analysis demonstrated good internal consistency for the scale.
• Construct validity was found to be acceptable, supporting the scale's measurement properties.

Conclusions:

• A validated learner-centered scale for AI tutoring acceptance quality was successfully developed.
• The scale offers a practical tool for evaluating AI-supported educational designs.
• This instrument can facilitate future intervention research in AI-assisted teaching contexts.