«NETWORKING IN EVERYDAY LIFE by Bernard J. Hogan A thesis submitted in conformity with the requirements for the degree of Doctor of Philosophy ...»
Individuals are rarely offered limitless opportunities to network with whoever they want and whenever they want. Rather, individuals maintain a relatively speciﬁc set of alters. Many of these individuals can be elicited through personal network capture techniques such as the name generator. Moreover, in addition to knowing the number of alters in the network and proportion of alters who are one type rather than another, one can consider the structure of the network itself. How individuals network may depend strongly on both the kinds of people in the network, but also how these people relate to each other in the eyes of the respondent. Chapter 6 examines this structuring of ties to illustrate how different social roles work either as bridges or bonders within the personal network and whether variations in the composition of the network are associated with variations in the ways in which individuals network with each other.
Research Question 4: What roles most evidently show a group-like structure across networks? Group-like structures in the network are relevant in several ways.
First, they indicate cognitive categories as much as actual social structures. That is, people may consider certain alters as belonging to a group whether or not these individuals actually possess this objective structure (Freeman, Freeman, and Michaelson, 1989; Freeman, 1992). Second, participants could indicate groups as a part of the socila network collection method. the process used to collect the social networks from participants included the ability for participants to signal groups, yet most network
of edges and work with the edges. By giving this matter scrutiny, I can talk about later models with greater conﬁdence. Third, Wellman’s persuasive theory of networked individualism frequently refers to a shift from groups to networks in social life. I use this theory in several areas of this work, and so it is relevant to understand who these groups are.
Research Question 5: How does the variation in contact with network members relate to structure? Taken solely as structures of relationships, social networks are clearly ordered, and roles help to deﬁne this order. By positing networked individualism as a theory of network composition, we can classify some roles as being more networked individualistic by virtue of their linking patterns (e.g., spanning the network rather than linking inwards towards homophilous roles). However, this is merely a secondary task (here, at least) to an appreciation of the networks as points of access for media use.
One premise of the theory of networked individualism is that networked individuals will make use of new media technology to facilitate more person-to-person interaction and less place-to-place interaction. Hence, media use will not superimpose itself cleanly over groups, but rather cut across speciﬁc boundaries, as individuals ﬁnetune their networking patterns with others (Kim, Kim, Park, and Rice, 2007; Wellman and Frank, 2001). This theory works well within an analysis of the social accessibility with alters. Namely, it suggests that access (both one’s ability to access alters and one’s actual behaviour) has changed as a result of the introduction of new media—or more precisely, that new media affordances facilitate new ways of interacting with alters enabling new patterns that are person/dyad speciﬁc rather than globally applied (Wellman et al., 2003).
In Chapter 6 I explore the variations in overall contact with the network (i.e. how frequently do individuals contact anyone). I investigate the extent to which these vari
networked individualism, there will not merely be person-to-person networking and speciﬁc networked individualistic structures, but the two should be associated with each other.
3.4 The alter level: Who gets access and why?
Regardless of the relationship between overall network structure and the variations in networking with others, there are still reasons why one might vary contact with speciﬁc alters. Saying this another way, Chapter 6 examines variability of media use with one’s network in general. Chapter 7 examines variation in the number of media used with speciﬁc network members controlling for network structure.
A prevailing hypothesis about this process is Haythornthwaite and Wellman’s theory of media multiplexity (1998). This hypothesis suggests that the more strongly tied ego is to alter, the more media ego will use with alter. However, the overall theoretical arc of this dissertation is that networking is about accessibility rather than tie strength.
Strong ties are usually accessible to some extent, but do people use more media with strong ties regardless of their social accessibility, or do people use more media with accessible alters regardless of closeness1. Recall that accessibility was deﬁned in Chapter 2 as an individual’s capacity to learn of or react towards the social activities of other individuals in such a way that it preserves or strengthens the pre-existing relationship or is considered non-disruptive. Using this deﬁnition, alters that are more spatially proximate are more socially accessible, as it involves less of a disruption to access them. Also, alters that are structurally embedded (Moody and White, 2003; Wellman, Carrington, and Hall, 1988), or share many mutual relationships with ego are more
3.5 Presenting a coherent framework How can these disparate questions from different levels of analysis be uniﬁed? In the conclusion I offer a story of multiple media use in everyday life that seeks to do just that. After having examined variations at these levels, I conclude that many differences in networking can be understood primarily as either individual attributes or propensities or matters of the speciﬁc relationship between ego and alter. However, because of these two phenomena, networking can be an increasingly differentiated affair. The norms for networking that are associated with small bounded communities are being replaced by person-speciﬁc gatekeeping practices (i.e. differences in social accessibility). Ultimately, the use of additional media may enable access to more alters or more diverse alters, but it also complicates the process of networking in general. Thus I conclude by addressing the conundrum: How is it that the use of multiple media simultaneously makes it easier to access individuals and yet more difﬁcult to network in general?
Chapter 4 Research methods and measures
4.1 Introduction data for this analysis comes from the Connected Lives Project, a study of
communication and media use in East York, Ontario, Canada. This chapter will present an overview of the research site as well as the speciﬁc methods used for capturing data. This includes a report on the composition of the data and its representativeness, an overview of the multiple instruments used to assess network size and structure as well as the instruments used to assess media use.
4.2 The research site East York is a 22.26 km2 area of Toronto just north east of the downtown core. Up until 1998, East York was a semi-autonomous borough of Toronto. In that year, the borough was dissolved as a separate governing area. Nevertheless, East York still maintains some semblance of a separate identity, with summer festivals, custom street signs, its own newspaper (The East York-Riverdale Mirror) and several community centres.
East York was selected for the research site by principal investigator Barry Wellman. Wellman had previously studied East York on two occasions. Initially, he was a
co-investigator of a research team studying interpersonal relations, mental health and social support in 1969. This study was followed up ten years later with Wellman’s second, longitudinal study of social networks and community. Both studies were considered particularly fecund, having spawned numerous papers including one each in the American Journal of Sociology (Wellman, 1979; Wellman and Wortley, 1990). In the original study design, Wellman suggests that East York was particularly appropriate for an analysis of community given its relative ethnic and cultural homogeneity. He was especially interested in the idea that a person’s set of intimate ties was only partially tethered to the neighbourhood, and actually extended out through the city and the world. Whereas the examination of a cosmopolitan and transient area might bias these results, he believed that the ethnic and cultural homogeneity of East York would be a more persuasive setting. The results of this analysis did indeed show that one’s personal ties were far ﬂung and disparate (Mok, Wellman, and Carrasco, 2008).
As a consequence of Canada’s immigration policy as well as geographically concentrated chain migration, the cultural make-up of East York has changed dramatically in the last 30 years. That said, this situation is far from anomalous. Rather, it is very representative of the changing ethnic landscape of Canada’s urban centres (Statistics Canada, 2003). As of 2001, over 40 percent of Toronto’s population were foreign born. While the changing cultural composition of East York is not directly addressed in this dissertation, it does help to colour and interpret the networking styles of the participants. Twenty-four percent of the survey and 17 percent of the interview are immigrants who have been in Canada less than ten years. As will be discussed in the results chapters, these individuals are especially prone to using ICTs such as webcams and online chat rooms as a means for keeping contact with their non-local
4.3 The survey implementation process The initial survey design called for a self-administered survey. The resulting document was a 32-page questionnaire. The questions were developed by the Connected Lives research team, a collection of 8 graduate students under the direction of Barry Wellman.1 The design and layout of the survey was done by myself using InDesign CS, an industry standard typesetting program. The survey is included here in Appendix A.
The sampling frame for the drop-off and pick-up process was a list of 1000 names and corresponding addresses stratiﬁed by FSA (Forward Sorting Area, i.e., the ﬁrst three letters of the postal address). Each name was sent an introductory letter letting them know that someone would be by with a survey in the forthcoming weeks. A team of about 20 undergraduate and graduate students performed the pick-up and drop-off process. The survey delivery process began in August of 2004 and continued until January of 2005. The status of each drop-off was recorded using an online database designed by myself and Michelle Levesque, an undergraduate computer science student. If a survey was not completed in two weeks, we mailed a reminder to the participants. After four weeks we mailed a second copy of the questionnaire to the respondents along with a self-addressed stamped envelope. After three unsuccessful attempts, the survey was abandoned. Upon completion of the survey respondents were given a 5 dollar coupon booklet for Tim Horton’s (a popular Canadian coffee chain) as a thank you gift.
The survey delivery process resulted in 350 completed and returned questionnaires. As only 621 of the 1000 names were valid, this results in a response rate of
as lack of language skills, frailty, death or having moved.
4.4 The interview implementation process Understanding the limitations of survey data, we designed a follow-up interview on the topics included in the Connected Lives survey. The interviews were designed by the same team of graduate students under Barry Wellman. Like the survey, this interview covered media use, social engagement, social support and household relations.
It also included a novel network analysis tool which we could not include in the survey due to its complexity. This instrument is discussed in Section 4.7.5.
The sampling frame for the interviews was taken from the completed surveys.
Every completed survey included a contract which was a single sheet kept separate from the survey booklets. In addition to reviewing the usual norms of survey delivery (such as conﬁdentiality, and how we would use the data), it also gave the respondents a place to indicate whether we could contact them for a follow-up interview. We gave them three options: yes, no, and maybe. We contacted everyone who answered yes or maybe. This was 170 of the 350 respondents. We contacted these 170 individuals, offering 20 dollars for a two hour interview.
Between January and April of 2005, 87 interviews were completed. Eighty-six of these are valid interviews for the analysis network herein, and one is excluded because no network analysis was performed (due to the respondent’s frail condition during the interview). Despite the fact that the interviews only represent 25 percent of the total
4.5 Representativity and basic demographics This section will compare some demographics from the 2001 Canadian census and the survey and interview samples. It will be shown that both the survey and the interview are marginally biased on several characteristics, but overall show a reasonable level of comparability with the target population. This means that it will be difﬁcult to assert point estimates about the population with absolute precision. However, given that this sample still represents a large, mainly representative swath of East York’s population, I should still be able to make reasonable claims about the correlations between speciﬁc variables without fear that the results are spurious due to sample bias.
Data from the 2001 Census was provided through Statistics Canada’s online community proﬁles.2 These data along with the corresponding percentages from the Connected Lives survey and interview are summarized in Tables 4.1, 4.2 and, 4.3. The ﬁrst table reviews personal characteristics (age, gender, and nationality). The second table reviews household characteristics (marital status, and family composition). The third table reviews socioeconomic indicators (employment status, income, and education).