The Incidental Pundit: Who Talks Politics with Whom, and Why?

View Researcher's Other Codes

Disclaimer: The provided code links for this paper are external links. Science Nest has no responsibility for the accuracy, legality or content of these links. Also, by downloading this code(s), you agree to comply with the terms of use as set out by the author(s) of the code(s).

Authors William Minozzi, Hyunjin Song, David Lazer, Michael Neblo, Katherine Ognyanova
Journal/Conference Name AMERICAN JOURNAL OF POLITICAL SCIENCE
Paper Category
Paper Abstract Informal discussion plays a crucial role in democracy, yet much of its value depends on diversity. We describe two models of political discussion. The purposive model holds that people typically select discussants who are knowledgeable and politically similar to them. The incidental model suggests that people talk politics for mostly idiosyncratic reasons, as byproducts of non-political social processes. To adjudicate between these accounts, we draw on a unique, multi-site, panel dataset of whole networks, with information about many social relationships, attitudes, and demographics. This evidence permits a stronger foundation for inferences than more common egocentric methods. We nd that incidental processes shape discussion networks much more powerfully than purposive ones. Respondents tended to report discussants with whom they share other relationships and characteristics, rather than based on expertise or political similarity, suggesting that stimulating discussion outside of echo chambers may be easier than previously thought. Acknowledgements. We thank Greg Caldeira, Jon Kingzette, Philip Leifeld, and Stefan Wojcik for advice and comments. A previous version of this paper was presented at the 2016 annual meeting of the Midwest Political Science Association. ∗Associate Professor, Department of Political Science, The Ohio State University. minozzi.1@osu.edu. Corresponding author. †Assistant Professor, Department of Communication, University of Vienna. hyunjin.song@univie.ac.at ‡Distinguished Professor, Department of Political Science and College of Computer and Information Science, Northeastern University, d.lazer@northeastern.edu §Associate Professor, Department of Political Science, The Ohio State University. neblo.1@osu.edu ¶Assistant Professor, Department of Communication, School of Communication and Information, Rutgers University. katya@ognyanova.net Forthcoming in the American Journal of Political Science Introduction Informal political discussion plays a fundamental role in modern democracy (Delli Carpini, Cook and Jacobs 2004), with its own set of normative criteria. Deliberative democratic theorists, in particular, emphasize the importance of “deliberativeness” in everyday political talk (Mansbridge 1999; Conover, Searing and Crewe 2002). Political conversation need not have the same qualities as formal, rule-bound deliberation to count as “deliberative” (Schudson 1997; Neblo 2015), but theorists suggest minimal criteria, including considering a variety of perspectives and alternatives. Therefore, if political conversation is to play its full, deliberative part in democracy, people must engage across di erences (Mutz 2006). If people choose discussion partners based on political similarities, however, they forgo those bene ts of discussion across di erence (Huckfeldt, Mendez and Osborn 2004). Some scholars suggest that political talk may be more likely between those with partisan or ideological similarities (Bello and Rolfe 2014); others that most people expose themselves to disagreement in everyday interactions (Huckfeldt and Sprague 1995; Mutz and Mondak 2006). But the etiology of this diversity—or its absence—remains unclear. Political talk is not necessarily driven by an intentional process. We describe two models from previous scholarship: a purposive model and an incidental model. The rst conceptualizes political conversation as goal-oriented. Citizens may be motivated to discuss politics to gather information that will help solve a problem, such as learning which candidate is best. In contrast, the second model posits that political discussion typically occurs as a byproduct of other social interactions. Consequently, whatever forces structure a citizen’s social environment also shape political conversations—if one’s social context is homogeneous, one’s political discussions will be too. Both processes generate homogeneity in political conversation. The motivations matter because they suggest di erent prospects for fostering cross-cutting political discussion for example, by encouraging participation in deliberative forums. We seek to identify whether political talk is more consistent with a purposive process or an incidental one. Though debate persists, the balance of evidence presented heretofore o er 1 more support for the incidental model. Assessing the relative importance of each, however, is challenging. Much prior work relies on egocentric surveys based on name-generator and nameinterpreter techniques (Klofstad, McClurg and Rolfe 2009). Such designs cannot easily account for the opportunities for conversation, thus confounding choice with opportunity. Recent work has yielded improved inferences by using more expansive name generators (Eveland, Appiah and Beck 2018) and even cross-sectional, whole-network data (Song 2015; Pietryka et al. 2018). However, inferences based on cross-sectional data or even a single, whole network have di culty ruling out alternative explanations that might produce similar patterns (Fowler et al. 2011). We overcome these inferential problems by analyzing a unique dataset: the Friends for Life Study. This project presents a whole-network, multiplex, panel dataset consisting of more than 100 networks from fourteen sites across the U.S. between 2008 and 2016. We surveyed these whole networks multiple times per year, with roster-style network batteries measuring several social relationships, including political discussion. Based on the Friends for Life Study, we nd strong, clear evidence that political talk is more consistent with an incidental process than a purposive one. Using out-of-sample predictive accuracy, exponential random graph models, and Bayesian hierarchical models to pool results from many networks, we show that political attitudes and identities are poor predictors of who talks politics with whom. We nd no evidence that individuals are more likely to engage with more interested or knowledgeable peers. Rather, in keeping with the incidental model, social relationships such as friendship predict discussion with much more accuracy. Furthermore, we nd no evidence of more subtle purposive motivations, e.g., that friends might be more more sensitive to di erences in political identities than acquaintances, avoiding politics to maintain relationships. Study participants appear neither to aggressively seek like-minded interlocutors, nor actively resist discussion across di erence. Consequently, we conclude that people may not cultivate homogeneous political discussion networks so much as happen into them incidentally.
Date of publication 2019
Code Programming Language R
Comment

Copyright Researcher 2022