Using Artificial Intelligence (AI) To Generate a Survey: A Step-by-Step Guide

• In Event: Survey Science in Education: Methods, Analysis, and Emerging Technologies

Abstract

Application of artificial intelligence (AI) has accelerated the survey development process and reduced the resource assumption. However, there is a lack of methodological guidance on how to develop a survey using AI. This study provided a step-by-step tutorial for survey construction with Automatic Item Generation and Validation via Network-Integrated Evaluation (AI-GENIE). The following five steps were suggested: (1) domain conceptualization and identification, (2) item pool generation and prompt engineering, (3) item pool reduction, (4) scale validation, (5) human expert panel review of items for content relevance. The proposed guidelines highlighted the importance of human-AI collaboration to ensure that AI-assisted scale development maintains both theoretical coherence and empirical rigor. Researchers may apply the guidelines in their empirical research involving scale development.

Authors

• Ruiqin Gao, University of South Carolina

• Christine DiStefano, University of South Carolina

• Xiaobo Wei, University of South Carolina

• Fang Wang, University of South Carolina