Can generative AI craft scale items? A mixed-method study on AI's capability to adapt and create new scales with recommendations for best practices

Abstract

This study explored the potential of generative artificial intelligence (AI), specifically Generative Pre-trained Transformer 4 (GPT-4), in revolutionizing the development and adaptation of research scales, a process traditionally characterized by its time-intensive nature and susceptibility to human bias. This study employed a mixed-method approach and assessed GPT-4's ability to generate reliable and valid scales and compared them with conventional methods. GPT-4 was chosen for its accessibility and accuracy in producing contextually appropriate items. Quantitative analysis confirmed the reliability and validity of the AI-generated scales, while expert evaluations highlighted the model's ability to enhance methodological efficiency and reduce bias. However, this study underscores the importance of combining AI's computational strengths with human expertise to ensure optimal outcomes. This research integrates generative AI into scale development and introduces innovative practices that streamline the process without compromising quality. These findings pave the way for future explorations of AI's role in research methodologies, and offer recommendations for best practices in using AI for scale development. This study contributes to the broader discourse on AI's transformative potential in research, and emphasizes its capacity to enhance the precision and efficiency of data collection tools.

Ähnliche Arbeiten