Corporate Social Responsibility Activities Through Twitter: From Topic Model Analysis to Indexes Measuring Communication Characteristics

The communication of corporate social responsibility (CSR) highlights the behavior of the business toward CSR and their framework of sustainable development (SD), thus helping policymakers understand the role businesses play with respect to the 2030 Agenda. Despite its importance, this is still a relatively underexamined and emerging topic. In our paper, we focus on what businesses communicate about CSR through social media and how this relates to the Sustainable Development Goals (SDGs). We identified the topics discussed on Twitter, their evolution over time, and the differences across sectors. We applied the structural topic model (STM) algorithm, which allowed us to estimate the model, including document-level metadata (time and sector). This model proved to be a powerful tool for topic detection and the estimation of the effects of time and sector on the discussion proportion of the topics. Indeed, we found that the topics were well identified overall, and the model allowed catching signals from the data. We derived CSR communication indexes directly from the topic model (TM) results and propose the use of dissimilarity and homogeneity indexes to describe the communication mix and highlight differences and identify clusters.