Academic Commons

Theses Doctoral

Networks in Polarized Times: What Americans Talk about, with Whom and When?

Lee, Byungkyu

What do American people talk about, with whom, and when? What people talk about influences and is influenced by who they talk with, and when they talk structures the interdependency between discussion topics and relationships. This context-alter-topic interdependency provides an opportunity to identify contextual mechanisms by which social and political networks are activated and deactivated in response to salient social events and polarized contexts. Do people talk about important matters with fewer people than ever before? Do people organize their political belief systems in ideological ways using the single liberal-conservative dimension? Do people discuss politics with more people who are more politically diverse in contested elections? This dissertation answers these and related questions by revisiting the same survey data that others have found useful in the past, but with a new and fresh lens.
Do people talk about important matters with fewer people than ever before? The 2004 General Social Survey (GSS) reported significant increases in social isolation and significant decreases in ego-network size relative to previous periods. These results have been repeatedly challenged, though none precisely identify the cause of decreased ego-network size. The second chapter shows that it matters that the 2004 GSS -- unlike other GSS surveys -- was fielded during a highly polarized election period. I show that political priming induced by presidential election events makes people frame "important matters" as political matters, and political polarization further suppresses network size especially for non-partisans.
Do people organize their political belief system in ideological ways using the single liberal-conservative dimension? By considering a set of interrelated political beliefs as a network of belief systems, the third chapter seeks to resolve theoretical puzzles concerning the organization of political belief systems, and address competing accounts of the role of political ideology and core values. I compare results from the American National Election Studies and General Social Surveys, and show the strong contextual influence on belief systems. I find that belief systems, often thought to be relatively stable, need to be "activated" by certain social cues.
Do people discuss politics with more people who are more politically diverse during contested elections? The fourth chapter focuses on battleground states to investigate the mutual interrelationship between political discussion partners and topics: who people discuss politics with depends on which issues they discuss and vice versa. I propose that increases in the salience of politics and exposure to opposing views contribute to the activation of interpersonal political echo chambers. I present evidence to support this claim based on statistical analysis of the 1992, 2000, and 2008 National Election Studies.
My dissertation shows, throughout three empirical chapters, why we need to seriously consider socio-temporal context in studies of social and political networks. I use survey timing, exogenous events, and battleground states to show how political situations induced by political events activate ideological thinking, which in turn deactivates our core discussion networks, and ultimately activates interpersonal political echo chambers. In sum, I discover situational activation of network processes.


  • thumnail for Lee_columbia_0054D_14587.pdf Lee_columbia_0054D_14587.pdf application/pdf 1.06 MB Download File

More About This Work

Academic Units
Thesis Advisors
Bearman, Peter Shawn
Ph.D., Columbia University
Published Here
May 14, 2018