The Productivity Paradox in Turkish Agriculture: Can AI-Supported Thematic Analysis Provide a Solution?/ Türk Tarımında Verimlilik Paradoksu: Yapay Zekâ Destekli Tematik Analiz Bir Çözüm Sağlayabilir mi?

Alper Tunga Alkan; Hakan Acer

doi:10.47105/nsb.1825526

The Productivity Paradox in Turkish Agriculture: Can AI-Supported Thematic Analysis Provide a Solution?/ Türk Tarımında Verimlilik Paradoksu: Yapay Zekâ Destekli Tematik Analiz Bir Çözüm Sağlayabilir mi?

Authors : Alper Tunga Alkan, Hakan Acer

Pages : 152-174

Doi:10.47105/nsb.1825526

View : 158 | Download : 259

Publication Date : 2025-12-27

Article Type : Research Paper

Abstract :This study addresses the time cost problem, one of the most significant limitations of thematic analysis among qualitative research methods, and proposes an AI-supported hybrid model to address this issue. The focus of the research is the “productivity paradox” (low productivity despite abundant resources) in the Turkish agricultural sector. In the study, Braun and Clarke’s (2006) traditional six-step thematic analysis protocol and Keshav’s (2007) three-pass systematic reading model were integrated with AI to develop a methodological framework. A total of 985 academic articles collected from the Web of Science (WoS) database using the keywords “Turkey” and “agriculture” were analyzed using two different large language models (LLMs): ChatGPT and DeepSeek. In the first stage, the 268 articles flagged by both models as “high relevance” were identified as the intersection set, and in-depth analysis was conducted on this dataset. In the second and third stages, AI capabilities such as code suggestion, thematic clustering, and consistency analysis were utilized; however, the final interpretation, synthesis, and meaning processes were carried out by human researchers. The findings revealed that the productivity paradox in Turkish agriculture stems from interconnected multidimensional themes such as natural resource management, technological adaptation, climate change, policy governance, socio-economic dynamics, and water management. It was observed that the proposed hybrid model accelerated and scaled the analysis process while maintaining methodological integrity by preserving the depth of human interpretation. This study demonstrates the potential of AI-supported thematic analysis to offer an innovative and efficient methodological framework for qualitative research in the social sciences.
Keywords : Türk tarımı, tematik analiz, yapay zekâ, hibrit model, nitel araştırma

ORIGINAL ARTICLE URL