Using Generative AI to Enhance Psychometric Scale Development in Market Research

Psychometric scales are foundational to market research, yet their development is traditionally resource-intensive, slow, and prone to cognitive and cultural bias. Item generation, the stage where abstract constructs are translated into candidate measures, is especially demanding. This paper introduces a methodological framework for integrating generative artificial intelligence (AI) into item generation, illustrated through a proof-of-concept study on the entrepreneurial mindset. We outline a structured human–AI workflow that combines large language model outputs with critical researcher oversight at key decision points. The process demonstrates how AI can rapidly expand item pools, reduce redundancy, and surface alternative phrasings, while expert validation safeguards theoretical fidelity and conceptual clarity. Our framework is both rigorous and adaptable. It provides detailed replication guidance, highlights ethical and practical considerations, and identifies how researchers can transparently document AI use. By showing how AI-assisted item generation can enhance efficiency without compromising validity, this paper rethinks scale development methods for an era where market research must be both conceptually robust and responsive to fast-changing contexts.