The Application and Comparison of Artificial Intelligence LLMs in Psychological Statistical Analysis

With the rapid development of LLMs (LLMs) in the field of natural language processing, there is an increasing interest in their applications within psychological research. This study focuses on three LLMs: OpenAI's GPT-4, Zhipu AI's GLM-4, and Alibaba Cloud's Qwen2.5, to examine their performance and potential value in psychological research. The analysis emphasizes five dimensions of the models' capabilities in handling text data: convenience, lexical statistical abilities, core topic extraction, sentiment analysis, and code generation support. Results indicate that GPT-4 excels in core topic extraction and programming assistance but falls short in convenience for file uploads and sentiment quantification. GLM-4 shows strong performance in sentiment analysis quantification but has limitations in text analysis, code generation, and overall balance. Qwen2.5 demonstrates robust capabilities in text understanding and sentiment classification, highlighting its practicality in processing Chinese text. The findings suggest that LLMs possess significant application potential in psychological research, providing effective support for in-depth cognitive and cultural studies while enhancing the efficiency and accuracy of psychological analysis.