Worth the Weight: Framing Social Processes as Weighted Social Networks
Open Access
- Author:
- Mc Millan, Cassandra Leigh
- Graduate Program:
- Sociology
- Degree:
- Doctor of Philosophy
- Document Type:
- Dissertation
- Date of Defense:
- May 05, 2020
- Committee Members:
- Diane Helen Felmlee, Dissertation Advisor/Co-Advisor
Diane Helen Felmlee, Committee Chair/Co-Chair
Steven Andrew Haas, Committee Member
Jennifer Lynne Van Hook, Committee Member
Bruce A Desmarais, Jr., Outside Member
Eric P Baumer, Program Head/Chair - Keywords:
- Social Networks
Weighted Networks
Homophily
ERGMs
International Migration
Adolescent Friendship - Abstract:
- The current dissertation focuses on weighted networks, or networks where individual relationships are assigned quantitative measures of tie strength. Most social network research treats social ties as binary variables in which relationships are either present or absent. However, weighted network data is collected frequently, especially when considering how our social connections relate to processes of inequality. In the current project, I demonstrate that it can be misleading to overlook the variability that defines tie weights when studying various social phenomena of interest to social scientists. In my first empirical chapter, I begin by exploring the theoretical and methodological puzzles that arise when studying homophily in weighted social networks. I define two variants of homophily that can arise in weighted networks: (1) strong tie homophily, or the tendency for ties with high values to cluster together similar peers, and (2) weak tie homophily, or the tendency for ties with low edge weights to connect same-attribute actors. To distinguish between homophily occurring among strong versus weak ties, I develop a new methodological technique that can be incorporated into an Exponential Random Graph Model (ERGM) framework and find that many networks are apt to be defined by observable differences in the magnitude of strong versus weak tie homophily. In subsequent empirical chapters, I apply these new methodological techniques to study social processes that operate at both the macro- and micro-levels. My second empirical chapter argues that through the application of a social network perspective, we can better conceptualize patterns of international migration and understand what social processes drive this phenomenon on a global scale. I begin by conceptualizing the phenomenon of international migration as a social network, in which countries are individual actors and ties are migration flows that are weighted by the number of people making an international move. Using data published from the United Nations on migration between 173 countries from 2010 to 2015, I apply valued ERGMs to disentangle the theoretically-motivated factors that drive international migration. While I find modest support that economic factors govern the patterns of international migration, I uncover convincing evidence that social factors and patterns of historical inequality can be generalized to explain patterns of migration of a global scale. My final empirical chapter focuses on how the patterns of adolescent friendship networks are shaped by individuals’ racial/ethnic backgrounds and gender, as well as broader social contexts. First, I ask whether the tendencies towards forming same-race and same-gender friendships are associated with the strength of ties connecting these peers. Then, I test whether these associations vary according to a network’s racial and gender distribution. I find that in non-diverse, primarily white networks, weaker friendship ties are more likely to connect same-race peers, while racial homophily is not associated with the formation of stronger friendship ties. However, as racial diversity increases at the network level, stronger ties become increasingly more likely to connect same-race peers, while the tendency for racial homophily to guide the formation of weaker ties dissipates. I do not uncover evidence that patterns of gender homophily vary across strong versus weak adolescent friendship ties. Overall, this chapter highlights the empirical value in simultaneously considering patterns of strong and weak tie homophily for certain individual-level characteristics of interest.