AI-empowered Scale Development: Testing the Potential of ChatGPT

S. Hoffmann, W. Lasarov () and Y. Dwivedi ()
Additional contact information
W. Lasarov: Audencia Business School

Post-Print from HAL

Abstract: AI-tools such as ChatGPT can assist researchers to improve the performance of the research process. This paper examines whether researchers could apply ChatGPT to develop and empirically validate new research scales. The study describes a process how to prompt ChatGPT to assist the scale development of a new construct, using the example of the construct of perceived value of ChatGPT-supported consumer behavior. The paper reports four main empirical studies (US: N = 148; Australia: N = 317; UK: N = 108; Germany: N = 51) that have been employed to validate the newly developed scale. The first study purifies the scale. The following studies confirm the adjusted factorial validity of the reduced scale. Although the empirical data imply a simplification of the initial multi-dimensional scale, the final three-dimensional operationalization is highly reliable and valid. The paper outlines the shortcomings and several critical notes to stimulate more research and discussion in this area.

Keywords: artificial intelligence; ChatGPT; ChatGPT-supported consumer behavior; scale development; validation (search for similar items in EconPapers)
Date: 2024-08
Note: View the original document on HAL open archive server: https://hal.science/hal-04607717v3
References: Add references at CitEc
Citations: View citations in EconPapers (1)

Published in Technological Forecasting and Social Change, 2024, 205, ⟨10.1016/j.techfore.2024.123488.⟩

Downloads: (external link)
https://hal.science/hal-04607717v3/document (application/pdf)

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:hal:journl:hal-04607717

DOI: 10.1016/j.techfore.2024.123488.

Access Statistics for this paper

More papers in Post-Print from HAL
Bibliographic data for series maintained by CCSD ().