Volume 12, Issue 3 (12-2025)
Human Information Interaction 2025, 12(3): 49-68 |
Back to browse issues page
Download citation:
BibTeX | RIS | EndNote | Medlars | ProCite | Reference Manager | RefWorks
Send citation to:
BibTeX | RIS | EndNote | Medlars | ProCite | Reference Manager | RefWorks
Send citation to:
Khondabi M, Zerehsaz M, Azimi A. Comparing ChatGPT-4 and Gemini 2.5 Pro in Meeting the Information Needs of Doctoral Students at Kharazmi University. Human Information Interaction 2025; 12 (3)
URL: http://hii.khu.ac.ir/article-1-3261-en.html
URL: http://hii.khu.ac.ir/article-1-3261-en.html
Kharazmi University
Abstract: (82 Views)
Purpose: This study aimed to compare ChatGPT-4 and Gemini 2.5 Pro in meeting the information needs of doctoral students in Persian. The focus was to examine how accurately, relevantly, clearly, practically, diversely, and effectively each model supports academic, specialized, and Persian-language information-seeking needs.
Method: The study used a mixed-methods approach. The participants included 36 doctoral students from the Faculty of Psychology and Educational Sciences at Kharazmi University during the 2024–2025 academic year, selected through purposive sampling. An initial questionnaire and a comparative questionnaire were used, and the students evaluated the relevant indicators after direct interaction with both models. In the qualitative phase, think-aloud protocols, semi-structured interviews, and observation of the interaction process using Morae software were employed. Quantitative data were analyzed using the Wilcoxon signed-rank test, while qualitative data were analyzed through coding in MAXQDA.
Findings: The findings showed that the superiority of the two models does not follow a uniform pattern. ChatGPT-4 was rated significantly higher in overall performance, and users found it more effective in initiating searches, organizing interaction, formulating the initial problem, and enhancing information-seeking confidence. In contrast, Gemini 2.5 Pro showed significant superiority in the accuracy of scientific and specialized responses, the diversity of information, the coverage of different aspects of the topic, and the practical usefulness of responses. No significant difference was found between the two models in terms of comprehensibility. The qualitative findings also indicated that Gemini 2.5 Pro was more strongly associated with content depth, comprehensiveness, and support for research work, whereas ChatGPT-4 was mainly described in terms of conversational fluency, response organization, and user reassurance.
Conclusion: The results indicate that the two models are complementary tools. ChatGPT-4 is better suited to initiating searches and organizing users’ thinking, while Gemini 2.5 Pro is more useful for specialized, research-oriented questions. However, neither model can replace academic databases, expert judgment, or the researcher’s critical evaluation.
Method: The study used a mixed-methods approach. The participants included 36 doctoral students from the Faculty of Psychology and Educational Sciences at Kharazmi University during the 2024–2025 academic year, selected through purposive sampling. An initial questionnaire and a comparative questionnaire were used, and the students evaluated the relevant indicators after direct interaction with both models. In the qualitative phase, think-aloud protocols, semi-structured interviews, and observation of the interaction process using Morae software were employed. Quantitative data were analyzed using the Wilcoxon signed-rank test, while qualitative data were analyzed through coding in MAXQDA.
Findings: The findings showed that the superiority of the two models does not follow a uniform pattern. ChatGPT-4 was rated significantly higher in overall performance, and users found it more effective in initiating searches, organizing interaction, formulating the initial problem, and enhancing information-seeking confidence. In contrast, Gemini 2.5 Pro showed significant superiority in the accuracy of scientific and specialized responses, the diversity of information, the coverage of different aspects of the topic, and the practical usefulness of responses. No significant difference was found between the two models in terms of comprehensibility. The qualitative findings also indicated that Gemini 2.5 Pro was more strongly associated with content depth, comprehensiveness, and support for research work, whereas ChatGPT-4 was mainly described in terms of conversational fluency, response organization, and user reassurance.
Conclusion: The results indicate that the two models are complementary tools. ChatGPT-4 is better suited to initiating searches and organizing users’ thinking, while Gemini 2.5 Pro is more useful for specialized, research-oriented questions. However, neither model can replace academic databases, expert judgment, or the researcher’s critical evaluation.
Keywords: ChatGPT-4, Gemini 2.5 Pro, information needs, information-seeking behavior, large language models, Persian, doctoral students, artificial intelligence literacy, Kharazmi University
Send email to the article author
| Rights and permissions | |
 | This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License. |
