A systematic approach to analysis of the text of social media

Ananda Mitra


In the five year period starting in 2005, one of the most significant and visible application of the Internet has been the exponential growth of people who have enrolled themselves with networking sites to establish virtual connections with people they might know in real life. This has been described in popular culture as “social media” and has certainly attracted significant popular media attention. This paper examines the phenomenon within the theoretical contexts of discursive and narrative construction of reality where individuals are able to use a variety of representation tools – from text to video – to create a personal narrative. Previously it has been suggested that the discourse on social media sites can be labeled as “narbs,” an abbreviation for “narrative bits.” The narb becomes the building block for the identity narrative that is discursively produced and circulated amongst those who are “friends” of the individual. Since the number of friends could be very large in the case of virtual connections, and not all might be considered friends in the conventional meaning of the term, it is becoming increasingly important to be careful about the way in which narbs are produced and used. There is growing evidence that institutions ranging from universities to law enforcement agencies are using the narbs to gauge the identity of individuals. The identity narratives then become central in making attributions about an individual, and eventually decisions about the individual are based on the attributions which themselves need to be problematized with respect to authenticity, trustworthiness and agency. This paper offers a framework of considering the narb in a careful and systematic way where observers interested in making critical decisions about individuals are able to make appropriate judgments about narbs.

Much of the analysis that is used at this time to make the decisions about individuals relies on a specific narb. For instance, a waitress at a popular pizza restaurant was terminated from her job for a single narb that was posted by her. In some cases applicants for jobs have been eliminated from consideration because of a single narb that someone else had created about the individual. Instances such as these point towards a need to examine the narbs about an individual in a more systematic manner. This paper suggests an approach where the examination of the narbs take into consideration several issues. First, it is important to consider the volume of narbs and their frequency to gauge if there is sufficient information to base judgments and if the information is recent and regularly updated. Second, it is important to consider the author of a narb and see who the “agent” is. Third, it is important to consider the content of the narb with respect to the way in which it is constructed using which specific representational tools such as text, images, video, etc. Finally, it is important to explore if the narb actually provides any spatial information about the individual and allows for understanding how an individual moves through space. Most of these are components of any narb and provide an understanding of each narb in these terms as well as an understanding of the narrative that is produced by a set of narbs. The authenticity and trustworthiness of the narrative is then dependant on the relative value of these different components of the narb. For instance, it would be unwise to attach a great deal of value to the identity narrative if the majority of the narbs are produced by someone other than the individual.

This paper will also argue that among all the four different components of the narb, the issue of agency is perhaps the most important. It is quite possible that an individual would have numerous narbs on their social media Web site, but if the individual also has a large number of friends then the larger number of narbs could well be a function of the size of the network, whereas most of the narbs are actually authored by someone other than the individual. A simple example of this phenomenon is the abundance of birthday greetings that might populate the profile of an individual on the date of birth where all the narbs are indeed authored by others who might have a tenuous real connection with the individual. Because of the way in which the number of connections can grow exponentially and the others can influence the identity narrative produced by narbs, it is particularly important to examine the agency of narbs before coming to specific conclusions about the nature of an individual.

It is important to consider a systematic approach to the analysis of narbs not only for the benefit of institutions that might rely on narbs to understand their people better, but also to allow individuals some basic guidelines for managing their narbs. There is some popular literature that offer “tips” about who to be-friend and what to say on social media sites. However, much of these are a-theoretical and somewhat inconsistent. It is important at this juncture to consider developing a system where there is a clearer understanding of the way in which narbs are constructed as discourse, and thus can be de-constructed to unravel the ethical components of the discourse. The qualitative aspect of the analysis can be well grounded in discourse and narrative analysis, but I would suggest in this paper that it is important to consider a numeric analysis of narbs before embarking on a discourse analysis to create the identity narrative which might eventually remain unreliable because of the nature of the narbs used for the analysis. Understanding the different components of a narbs as suggested here could help avert the pitfall of mistaken attributions based on unreliable narbs.