International Open Access Journal Platform

基于自然语言处理技术的心理学研究方法

Psychological Research Methods Based on Natural Language Processing Technology

Psychology of China / 2026,8(1): 31-35 / 2026-01-16 10 6

View Download PDF

• Authors: 吕嘉乐
• Information:

  广西师范大学教育学部心理学系，桂林

• Keywords:
  Natural language processing; Psychological measurement; Psychological constructs; Large language models
  自然语言处理; 心理测量; 心理构念; 大型语言模型
• Abstract: Recent advances in natural language processing (NLP) have provided psychology with novel methodological tools. Language, as a key carrier of human psychology and behavior, enables the automated revelation of psychological characteristics across individuals and groups, spanning dimensions such as emotion, cognition, personality, social relationships, and culture. This article systematically reviews the latest developments in the application of NLP within psychological research. It focuses on its use in psychological construct mining, multilingual text analysis, automatic scale item generation, machine-assisted hypothesis generation in social psychology, group-level psychometrics, and the prediction of mental health intervention outcomes. Research indicates that NLP methods based on large language models (LLMs, e.g., GPT, BERT) not only enhance the efficiency and objectivity of psychological measurement but also expand the diversity and ecological validity of research samples. However, these approaches still face challenges related to data representativeness, model bias, interpretability, and ethical privacy concerns. Future research should deepen interdisciplinary collaboration between psychology and computer science and advance the development of interpretable NLP techniques and cross-culturally adaptive models. 近年来，自然语言处理（NaturalLanguageProcessing，NLP）技术的快速发展为心理学研究提供了新的方法论工具。语言作为人类心理与行为的重要载体，其自动化分析能够揭示个体及群体在情感、认知、人格、社会关系与文化等多个维度的心理特征。本文系统梳理了NLP在心理学研究中的最新进展，重点探讨了其在心理构念挖掘、多语言文本分析、量表项目自动生成、社会心理学假设辅助生成、群体心理测量以及心理健康干预效果预测等方面的应用。研究表明，基于大型语言模型（如GPT、BERT）的NLP方法不仅提升了心理测量的效率与客观性，还拓展了研究样本的多样性与生态效度。然而，该方法仍面临数据代表性、模型偏见、可解释性及伦理隐私等方面的挑战。未来研究应深化心理学与计算机科学的跨学科合作，并推动可解释NLP技术与多文化适应模型的发展。
• DOI: https://doi.org/10.35534/pc.0801005 (registering DOI)
• Cite: 吕嘉乐. (2026). 基于自然语言处理技术的心理学研究方法. 中国心理学前沿, 8 (1), 31-35.