Literature Reviewing: Addressing the Jingle and Jangle Fallacies and Jungle Conundrum Using Graph Theory and NLP
Yuanyuan Song, Richard Thomas Watson, Zhao Wang Xia · Journal of the Association for Information Systems · 2021
Scientific advancements in all fields, including IS, are built on previous accomplishment. Identifying similar causal models is critical for synthesizing research. However, the growing knowledge repository and inconsistencies in existing literature (i.e., jingle and jangle fallacies) challenge humans’ bounded rationality. Humans need supporting information systems to make a jungle of causal models amenable to analysis. This paper proposes using graph theory and natural language processing (NLP) methods to analyze knowledge networks and report similarity scores for causal models. This method builds on the first phase of the Theory Research Exchange (T-Rex) project, in which guidance on digitizing the core knowledge in publications is established. Digitizing core knowledge will provide an efficiency gain as illustrated in this paper and be a significant step forward for the knowledge economy.