The exponential growth of “Big Data” has given rise to a field known as computational social science (CSS). The authors view CSS as the interdisciplinary investigation of society that takes advantage of the massive amount of data generated by individuals in a way that allows for abductive research designs. Moreover, CSS complicates the relationship between data and theory by opening the door for a more data-driven approach to social science. This chapter will demonstrate the utility of a CSS approach using examples from dynamic interaction modeling, machine learning, and network analysis to investigate organizational communication (OC). The chapter concludes by suggesting that lessons learned from OC's history can help deal with addressing several current issues related to CSS, including an audit culture, data collection ethics, transparency, and Big Data hubris.
TopIntroduction
Many have said we live in an information age, one that emphasizes an ever-increasing ability to retrieve and acquire a plethora of data from a vast amount of resources. Much of this discussion highlights individuals’ abilities to act as consumers of information or at least as beneficiaries of mostly large organizations such as Google or the U.S. government to acquire and process data. However, there has been less emphasis on how the information age allows individuals to act as producers, whether conscious or not, of information. Indeed, from every email, text message, financial transaction, Facebook status update, Tweet, and even keystroke, individuals are constantly producing information about themselves and others, and it is increasingly the case that they can analyze this data themselves.
The result is the accumulation of what is more commonly known as Big Data. The term refers to not only the size of data, but also its ability to take many different forms and speed at which it is generated (Zikopoulos & Eaton, 2011). Computational social science (CSS) emerged because of Big Data’s overall size and temporal nature and the frequent inability of traditional quantitative methods (e.g., general linear model) to analyze the complexity of Big Data (Lazer et al., 2009).
The purpose of this chapter is to define CSS and describe the potential of CSS for organizational communication (OC). The next section focuses on description by defining three features that underlie CSS and their philosophical assumptions. Then, there will be three brief examples of research related to OC involving methods such as relational event modeling, machine-learning classification, and descriptive network analysis. Finally, the conclusion describes current issues confronting CSS and how insights from OC can help address them.