国际中文开源期刊平台

基于ChatGPT 的中文版情绪稳定性问卷的开发探索与信效度验证

Exploration of the Development and Validation of the Reliability and Validity of the Chinese Emotional Stability Questionnaire based on Chat GPT

中国心理学前沿 / 2024,6(3)：328-336 / 2024-03-21 315 184

查看全文 下载文档

• 作者： 高垚杰     
• 单位：

  天津师范大学心理学部，天津

• 关键词： ChatGPT；问卷开发；情绪稳定性；信度；效度
• ChatGPT; Questionnaire development; Emotional stability; Reliability; Validity
• 摘要： 随着情绪稳定性越来越受到社会与研究者的广泛关注，并鉴于其对于国民健康及多个研究领域的重要作用，质量良好的情绪稳定性问卷显得尤为关键。但目前情绪稳定性的测量多借助于大五人格量表中的神经质维度，既不利于大众的理解与传播，也揭露了本土化测量工具有待丰富的现状，因此本研究旨在开发中文版的情绪稳定性问卷，并顺应人工智能的发展潮流，学习国外最新的项目开发范式，探索利用ChatGPT生成中文版情绪稳定性项目的可行性，并通过专家筛选、项目分析、信效度验证等步骤，验证GPT生成项目的性能，以期形成一个质量良好的中文版情绪稳定性问卷。
• As emotional stability has received more and more attention from society and researchers,and in view of its important role in national health and several research fields (e.g., organizationalmanagement, mental health, education, physical health, decision-making behavior, etc.), good-qualityemotional stability questionnaires have become particularly crucial. However, emotional stability asa personality trait is currently measured by the neuroticism dimension of the Big Five scales, which isnot conducive to public understanding and dissemination and also exposes the fact that the localizedmeasurement tools need to be enriched. Therefore, this study aims to develop a Chinese version of theEmotional Stability Questionnaire and, in line with the development trend of artificial intelligence,learns from the latest foreign paradigm of item development and explores the feasibility of using thelatest version of ChatGPT for automatic item generation to develop a Chinese version of the EmotionalStability Questionnaire. The formal administration of this study included 68 items generated by the GPTretained after expert screening and a total of 42 items measuring neuroticism from four widely used BigFive personality scales (CBF-PI-B, Chinese version of the BFI-2, translated TIPI-10 and IPIP-BFAS), witha total of 479 valid samples obtained. The performance of the GPT-generated items was verified throughthe testing steps of item analysis, structural validity and Criterion-Related Validity calculation, reliabilityanalysis, and comparisons, resulting in a good-quality Chinese version of the Emotional StabilityQuestionnaire.
• DOI： https://doi.org/10.35534/pc.0603037
• 引用： 高垚杰．基于ChatGPT的中文版情绪稳定性问卷的开发探索与信效度验证［J］．中国心理学前沿，2024，6（3）：328-336．