Akademik Veri Yönetim Sistemi
MBAA International Conference, Illinois, Amerika Birleşik Devletleri, 22 - 24 Mart 2023
This study attempts to summarize tax research over the course of the last century in a fully encompassing approach that ignores author bias. Typically, literature reviews of tax research explore specific areas, and these provide the advantage of a comprehensive investigation of a specific topic. However, they invariably focus on the interests of the author. In contrast, this study uses a methodology that is agnostic to tax area. It uses techniques that are now available through modern software and statistical methods, yet captures work done nearly 100 years ago. Our study uses text mining tools and natural language processing (NLP) to examine tax research topics from the years 1926 to 2021, comprising a total of 102,717 references, curated to 3920 in tax. The result is an expansive summary of the areas that have proven the most historically relevant to academic researchers. As such, it provides a large-picture view of accounting research in tax. We anticipate that our findings will be useful to academic researchers in contextualizing their work, policy makers in understanding areas of strength and weakness in supporting their endeavors, and practitioners in identifying constructive partnerships to form with academics. From an education standpoint, our paper can be used as a guide to doctoral students identifying promising areas for future research, including guidance on the publication potential in well-regarded journals. In addition, from a methodology standpoint, we suggest our work as a springboard for further literature reviews invoking our text mining approach.
Keywords: data mining, text mining, natural language processing, tax history, accounting history