Introduction to the thesis: an overview of what is to come
“This thesis will examine how social structure contributes to social influence over the Internet.”
Everything has been written once. I’m at base camp. Tomorrow, I start the climb to the peak. Hoping for an avalanche-free ascent, but I expect a few crevasses.
In the meantime, here’s an overview of what I’ve been working on for the past 4 years. I’d love any feedback you might have; this is the introduction chapter (draft 1). Does it make sense? I’ve lost all perspective.
It should be easier to read than the other content I’ve thrown up here.
This thesis will examine how social structure contributes to social influence over the Internet. There are three objectives: first, to establish both theoretically and empirically that online communities are media rich enough to support social influence; second, to test assess relationships between social network analytic and social psychological attributes that are theoretically linked with social influence; and third, to test the relative importance of each in attitude development and behaviour conversion.
The Internet is an important new platform for information sharing. Widespread, free and open information sources like weblogs and forums produce libraries of user-generated information that can be accessed and repurposed by consumers, challenging the traditional models of knowledge ownership and distribution via top-down sources like newspapers, radio and television. Influence theories that focus on notions of interpersonal trust (e.g., Bandura, 1977), source credibility and expertise (e.g., Hovland, 1953; Petty & Cacioppo, 1986) are potentially contested in the new web paradigm, where identity is fluid (Turkle, 1995) and rapid dissemination of new ideas and disinformation from anonymous sources is made possible because of the ease with which the digital content can be reproduced and shared.
Further, paradigms that emphasise social norms appear inappropriate to apply to a medium in which communication is so lean that significant groups would seem unlikely to form. Yet although early accounts constructed these environments as dehumanised, antisocial and lawless spaces (Kraut et al), extensive research over the past 20 years has documented the plethora social structures that have emerged from the widespread webs of interconnections between people gathered in communities of practice (Wellman, 1999, 2001, 2002; Haythornthwaite, 2007). Ethnographers, anthropologists, sociologists, economists, political scientists and psychologists have found extensive evidence for human systems of self-organisation (Correll, 1995), governance (Dibbell, 1996) and, indeed, influence (Guadagno & Cialdini, 2005). These outcomes have emerged from online spaces that encourage social interaction, from e-commerce sites where reviewers develop reputations (Rheingold, 1993) to chatrooms where new patterns of language are established (Boase & Wellman, 2006), to sprawling virtual worlds where economic value based on rarity and social capital allow for the construction of culture in disembodied cyber-spaces.
Research that has sought to understand online influence has tended to focus on experimental situations, using short-term computer-mediated systems to assess the impact of anonymity on conformity (e.g., Spears & Lea, 1991; Guadagno & Cialdini, 2005). This research has described how the Web creates the oft cited “levelling-ground”, at once devaluing offline status, but at the same time making salient offline commonalities that are wittingly or unwittingly communicated between participants. Results of these analyses have highlighted the benefits of such systems for stigmatised and low-status populations, but they also have cautioned that the deindividuation of online activity supports compliance, leading to pockets of homogenous groups of individuals seeking confirmation from others perceived as similar. Arguably, these outcomes do not take into account the online identities developed by participants invested in the social systems of online communities (Guadagno & Cialdini,2 005); this research seeks to address this.
Both social psychology and social network analysis have theories about what contributes to influence. Social psychological approaches like the Elaboration Likelihood Model of Persuasion (Petty & Cacioppo, 1986), Conflict-Elaboration Theory (Perez & Mugny, 1996) and Social Identity Theory (Tajfel & Turner, 1979) emphasise the characteristics of the source and the interpersonal context, and their impact is on the target. Social network analysts emphasise features of the social context, like the number of connections an individual has to people who have already adopted an innovation, and the structure of the interconnections both within local groups and across the whole system. These are complementary approaches to social influence that, when combined, may go some way in explaining attitudes and behaviours in online situations. However, psychological research rarely considers structure and position in the social network (Weenig, 1993; Weening & Midden, 1991), and social network analysts rarely consider why, psychologically, their measurements are effective predictors of social influence (Weenig, 1993). This thesis seeks to unpack the subjective meanings of the network measures for community participants, and to identify elements of network analysis that are not explained psychologically.
We aim to shed light on these questions by examining the psychological content of relationships in Second Life, an online community designed by its creators to be used for sociability, and which boasts a population in the millions with an active virtual economy. We will examine how attitudes cluster in the social system and how a behaviour diffuses around its account holders. Data is gathered in two ways: first, with online surveys designed to capture interpersonal connections and the subjective perceptions of those relationships from the points of view of the respondents; second, by accessing behaviour content from the computer servers.
The present thesis provides evidence that virtual communities support identity principles; influence may occur based on whether the offline self or the online social identity are salient, and how the individual negotiates the two. In other words, offline personal identity informs attitude and behaviour decisions within the context of how decisions will affect the social standing of the online personal identity in the virtual group. This research therefore supports the idea that online community can be a significant source of normative influence, with effects on the offline attitudes and behaviours of the individual.
It further provides evidence that social network analytic constructs are psychologically based; network integration acts in general as a proxy for perceptions of trust, credibility and expertise, and identifies individuals who are sources of social comparison and embody the norms of their groups. However, the influence of these network and psychological attributes depends on the expected social outcomes for the individual of adopting a behaviour or belief. Further, the research supports both structural and psychological explanations for influence; both offer unique contributions to attitudes and behaviours.
Overview of the chapters
Chapter 2 provides an introduction to influence theories in the social psychological and social network paradigms, investigating how each is represented offline. It seeks to identify the common processes and establishes the areas where structural and psychological explanations may diverge.
It continues with an analysis of the unique features of online communities that challenge the influence theories. As will be shown, the deindividuated nature of identity, the meanings ascribed to depersonalised, disproximate relationships and the development of online-only social identity based on a sense of belonging rather than physical place particularly challenge the identity approaches to influence within psychological thought.
Chapter 2 then outlines the ways psychologists and social network analysts have described influence in computer-mediated communication. However, overall, the studies about online influence in both disciplines have ignored the meanings of online communities for participants, whether by examining online influence with disinterested participants or by describing social structure using data-driven methodologies that do not take into account the participant’s experiences. This leads to the research questions and a discussion of how we intend to answer them.
Chapter 3 introduces the tools used to answer the research questions. First, it introduces the Second Life domain, describing the design features that encourage the development of personal identity, interpersonal relationships and social identity. It continues by outlining the methods used to collect data in ego-centric and socio-centric network studies, the two types of social network analysis undertaken in this research. For the former, sociometric surveys were developed to capture the meanings of relationships between people. For the latter, access to the Second Life servers was arranged with the support of the virtual community owners to capture the connections between all community members in the sample, plus the behavioural outcomes described in Study 3. The ethical implications are outlined, focussing on participant anonymity, secondary sources and informed consent.
Research in online environments requires special ethical considerations, and these are also discussed at the end of Chapter 3.
In Chapter 4, we explore the relationships of Second Life. This investigation seeks to understand the psychological nature of the relationships in this online community and to establish behavioural indicators for interpersonal closeness and distance. In addition, it identifies the relationships between network measures of cohesion and influential source attributes.
The data scrutinised in Chapter 4 support hypotheses about virtual community as a location for social influence. Members of this online community are found to have relationships based on trust and actively compare themselves with one another in situations where they are uncertain about social outcomes in their virtual groups. Participants are able to identify virtual sources of credibility and expertise and consensually pinpoint actors who represent the prototypes of their online groups. These analytic findings extend offline influence research by applying the principles to the online environment. In particular, they establish that online groups can engender a social identity that has the potential to influence.
Further, the findings of this research establish a strong relationship between social network analysis and social psychology; definitions of “closeness” from both disciplines appear to be proxies for one another. This finding helps to identify the best network measure for describing interpersonal connectivity that is most unique from psychological phenomena, used in answering the questions moving forward in the thesis.
In Chapter 5 we seek to understand how attitudes cluster in this online community. It explores whether psychological or network features better predict opinions about cybersex by measuring an ego-centric network using self-report online surveys to gather data about network connections and psychological attitudes.
Sexual activity in online environments like Second Life is commonplace, but not a practice all members of the community engage in. It is a phenomenon that primes both online and offline personal and social identities, bringing to bear questions of cultural morality and personal proclivity. In Chapter 5 we explore this duality which challenged participants particularly because the virtual medium offers little corroborating evidence for actual practices and stated beliefs. This leads to new insights into the effects of the looseness of Internet networks on the generation of norms, resonating with Wojcieszak’s (2005) notion that online pluralistic ignorance engenders confirmation biases that lead to peer pressures and greater extremism.
Chapter 6 uses behavioural and self-reported data to establish the interactions between the normative, network and interpersonal influences and personal attitudes. It describes the diffusion of a new innovation, voice communication, by observing the pathways of adoption through the whole Second Life network. It shows that attitudes and intentions predict uptake, that the implications of the innovation on the negotiation of online and offline identity affect uptake, and finally, that social psychological and social network features predict diffusion in unique ways.
Finally, Chapter 7 provides a summary of the findings and the conclusions drawn regarding the unique roles social psychological and social network attributes play in social influence. In particular, and with reference to the online context, this is discussed by looking at the negotiations of identity virtual community participants must make to maintain consistency and social acceptance; by looking at the effects of the structure of the online network on social norms; and by looking at the implications of belief and behaviour innovation content on the pathways of diffusion. These are contrasted with the assumptions made by the social psychological and social network models presented in Chapter 2.
The chapter also presents the methodological issues this study raises, in particular the best ways to isolate network and psychological measures, and how to study online networks. It concludes with suggestions for future research.