Social Network Construction in the Information Age: Views and Perspectives

Introduction

Generally speaking one's social circle, or more specifically one's social network, includes friends, family, colleagues and acquaintances. Social connectedness however, is a psychological term which describes the duration, frequency, familiarity and reciprocal nature of the relationships people have with others in this circle or network. Both our social network and how connected we feel to people within it are important aspects of one’s social wellbeing. Our social circle and how we perceive our degree of social connectedness can be subjective and possibly difficult to articulate.

Classically, social scientists interested in such social networks and connectedness would perform small scale data collection studies. Such data collection traditionally consisted of observation or manually conducting interviews and questionnaires on a population of interest, to derive a list of ties between the members of the population which can later be studied from a sociological perspective. Clearly, this is a time consuming and expensive process. In addition, the ability to document networks and our connectedness over the course of time is limited by the frequency of the data collection activities. The realisation that both human memory and our perception of social connectedness are limited implies that any data collected through such manual processes will inherently contain errors. Regardless of all these problems, such data collection methods have represented the gold standard for some time.

By contrast, consider the rapid expansion of technologies which collect data either explicitly or implicitly about our social networks and connectedness. Social networking services have become part of most people’s daily lives. These services support people’s connections and interactions while keeping an electronic trace of all these activities. There is also a wealth of electronic data that is available as a bi-product of other processes (sales, travel, phone call logs and email server logs). Within this data are the digital footprints of activity from which social networks and our connectedness can be deduced. In such cases the data is not collected with the explicit purpose of being studied from a social network perspective. This aspect shifts the design decisions on electronic data to a later processing stage once the data is already available, rather than the data collection stage before the data is collected. This shift introduces a different set of decisions and processes when dealing with electronic data collection.

When considering electronic data, from which we wish to extract both social networks and information on connectedness, there are a number of problems to consider. Firstly, records may relate to an entity relevant to the business (customer, caller, traveller, recipient). These entities must be resolved into the actors within our social network. Next, methods to relate the actors must be determined based on relevant context provided within the domain in question. For instance, actors may be linked in an online store if one sends another a gift or if two people travel together. Finally, it is important to differentiate between relationships of different strength. One cannot possibly have equal relationships to all 500 friends on a social networking site. In this case, identifying the strong and weak relationships is important for a better network understanding.

In this chapter we will discuss the differences between manual data collection and electronic data collection to understand the advantages and the challenges brought by electronic social network data. We will discuss in detail the processes that are used to transform electronic data to social network data and the procedures that can be used to validate the resultant social network.

The rest of this chapter is arranged as follows: We first give a brief overview of manual social network data collection to orient the reader who is not familiar with the general practices in this area. In the second section, we describe electronic data collections and give some examples of studies that make use of this kind of data. Following this, we compare the two types of data collection and then proceed to explain the processes involved in electronic data collection. As part of the process involved we describe the steps of actor identification, tie inference, tie strength measurement and network validation. We conclude this chapter by giving some recommendations on future research related to network construction.