Selecting In and Out of Politics (Pre-Recorded)

• In Event: Reconsidering Polarization

Abstract

As Americans come to associate the word “politics” with conflict, it changes the expectations people bring to information contexts deemed “political.” Rather than preparing for open-minded engagement, they ready themselves for a fight, and this influences the way they process information (Groenendyk and Krupnikov 2020). But, if the association between politics and conflict influences people’s actions within political contexts, might it also influence who selects into and out of those contexts in the first place?

To answer this question, we designed a series of experiments examining how the label of “politics” affects self-selection. In each study, participant were either told they would be asked about politics (Study 1), given an option to participate in a political survey versus a consumer survey (Studies 2 and 3), invited to participate in an online political focus group (Study 4), or invited to a fictitious dinner party in which politics would serve as the topic of discussion (Study 5). In each case, we found that the mere label of “politics” had large and significant effects on self-selection: On average, partisan and ideologically polarized conflict-seekers selected in, and ideologically moderate and independents averse to conflict selected out. When we changed the labeling to suggest “general public agreement” on the political issue being discussed (Study 2) or referred to the survey as a “political deliberation” study (Study 3), it did little to change these results.

We believe these findings have vital substantive and methodological implications. First, surveys (and survey experiments) that identify their purpose as related to “politics” may systematically overestimate polarization and incivility by leading less polarized and more conflict avoidant individual to select out of the sample. When overestimates of polarization are disseminated to the public, they create a feedback loop, strengthening the expectation conflict in politics and fuel self-selection bias in future studies. Second, as suggested most clearly in Study 5, this same process is likely playing out within the public sphere (see also Neblo Esterling, Kennedy, Lazer, and Sokhey 2010), motivating individuals seeking a fight to select into politics and leading individuals hoping to avoid conflict to select out.

Thus, our results suggest that, while polarization is a very real problem, it is being perpetuated by our perceptions more than substantive disagreement. This is consistent with the literature, which shows relatively little polarization over public policy (Fiorina, Abrams, and Pope 2010), yet substantial affective polarization (Iyengar, Sood, and Lelkes 2012). Due to the outsized attention lavished on the most extreme voices, Americans have come to hold ridiculously caricatured images of their political opponents (Ahler and Sood 2018). But not all Americans are affectively polarized. Many are turned off by polarization (Groenendyk, Sances, and Zhirkov 2020, Klar and Krupnikov 2016; Klar, Krupnikov, and Ryan 2019), and while exposure to polarized media leads people to hate the opposition, it can actually lead to moderation of one’s own views (Levendusky and Malhotra 2015). Unfortunately, between the lack of attention to this group, and the self-selection problem identified in this paper, their voices are not being heard.

Authors

• Eric Groenendyk, University of Memphis

• Yanna Krupnikov, Stony Brook University