Bibliometric and Topic Modeling Analysis of Corporate Social Irresponsibility

This study provides the first hybrid model by applying bibliometric analysis and topic modeling for mapping the intellectual structure of CSiR literature to improve business model flexibility. It has the following research objectives: (1) overview of CSiR literature for publications trends, influential studies, significant authors, dominant journals, and location of authors (literature evolution); (2) prevalent and emerging themes in CSiR (topic identification and labeling); (3) topics development in CSiR literature (topic development and trends); and (4) academia progress in CSiR research, and directions for future CSiR research. Drawing on documents selected from 1990 to mid-2021, we propose a hybrid model for our study: a mix of (1) bibliometric analysis through the VOSviewer program, wherein we use citation analysis, keywords co-occurrence analysis, co-citation analysis, and co-authorship analysis; and (2) topic modeling exercise through a machine learning approach. We also observe current and emerging topics by building a topic-word distribution matrix, analyze their development, and witness topic initiation across different periods. Further, we identify the details of academic progress in CSiR that impede the development of CSiR literature. Underdeveloped CSiR conceptualization and operationalization impede the growth of CSiR literature. We find that there is less collaboration among authors for CSiR research. Lastly, we conclude by presenting potential areas for future research.