In many friend groups, politics is not the most popular topic of discussion. The potential for tension puts a damper on informal political conversations which have valuable democracy-sustaining benefits. Nevertheless, mutual respect between peers may prevent the kinds of explosive disagreements commonly seen online, and those who share nonpolitical similarities may be more likely to bend towards more moderate stances.
But will this same openness translate to situations where politically different individuals remain strangers? And, if so, could this effect be significant enough to contribute to depolarization?
In a new study published in the Proceedings of the National Academy of Sciences and led by Penn Integrates Knowledge University Professor Duncan Watts in collaboration with colleagues from Heidelberg University, University of California, Santa Cruz, and Microsoft Research, the researchers explore how the mechanisms behind intra-group receptivity can be applied to anonymous, informal political communication. The study shows that, by fostering feelings of closeness through incidental similarities, those with strong beliefs on either end of the political spectrum can begin to converge upon more moderate views.
The effectiveness of non-political cross-partisan bridges
The coauthors used a large-scale, two-phase experiment to determine how individuals’ receptiveness to political arguments is influenced by feelings of closeness with the people presenting them.
In the first phase, participants were asked to answer a series of questions about their nonfocal attributes — or features unrelated to political issues — including demographics, hobbies, and personality. Immediately afterwards, they were asked a separate set of questions assessing their political leaning, perception of inequality in the U.S., and stance on wealth redistribution. Finally, they were asked to describe their views on the topic in their own words via a short essay.
About six months later, the second phase of the experiment used this information to match the respondents with a new set of participants. After administering the same nonfocal and stance-measurement surveys as in Phase 1, the authors randomly assigned every new participant to a Phase 1 participant with either high or low non-political similarity and high or low agreement on the focal issue. Each new participant then saw a profile page of their match highlighting all of their common answers to the nonfocal survey, and were asked to rate how connected they felt with their match on a seven-point scale. Finally, the Phase 2 participants read the essays written by their matches, and were once again asked about their own political stances and how they perceived their match.
The results of these updated survey responses revealed intriguing insights into the mechanics of political depolarization. As can be expected, Phase 2 participants with strong political views became more extreme when matched with same-stance Phase 1 respondents. Contrary to previous research, however, the team found that interactions between participants with differing stances depolarized strong views on both ends of the spectrum, with both anti- and pro-redistribution respondents reporting more moderate updated stances. This effect was strongest among participants who indicated feeling closer or more connected to their Phase 1 matches based on their profiles; every unit increase in closeness corresponded with a 16.3% increase in the odds of reducing the pair’s consensus gap.
The authors also examined the opposite process — how viewing a match’s political stances affected feelings of closeness — by differentiating between expected closeness (how close a participant felt toward their match before reading the political essay) and experienced closeness (how close they felt after reading the essay.) While they found that both measures were strong predictors of how Phase 2 participants updated their reported political stances, they were not necessarily consistent before and after interacting with a match; reading an essay from a like-minded person increased perceived closeness, while reading one from someone with opposing views dampened these feelings. Furthermore, these effects were not symmetrical, since closeness decreased much more after interacting with someone with opposite views than it increased after interacting with someone with similar views.
Towards reducing polarization?
This work sheds valuable light on the dynamics of political communication and how even political hardliners can be swayed, showing that increased feelings of closeness towards the source of a political message can foster openness towards opposing views—a finding which subverts previous research showing that such exchanges can cause people to simply dig in their heels.
With this in mind, this research has important implications for reducing political polarization on online media platforms. Much like in traditional friend groups, many people online prefer to keep politics outside of their personal social networks, and typically consume news and other political content which is consistent with their beliefs. Interventions designed to expose users to strangers with different views while highlighting nonpolitical affiliations could combat polarization, mimicking the kinds of informal opinion-sharing and receptiveness to ideas found in friend groups which are otherwise too politically homogenous to have depolarizing effects.
This work complements the CSSLab’s Penn Media Accountability Project (PennMAP), which aims to better understand the information ecosystem — and how it can foster harmful processes such as polarization — using large-scale cross-platform data. The study contributes to this mission by highlighting the interpersonal mechanisms behind openness to ideas, revealing new opportunities to create cross-partisan bridges.
Duncan Watts is the Stevens University Professor and a Penn Integrates Knowledge University Professor with appointments in the School of Engineering and Applied Science, the Annenberg School for Communication, and the Wharton School. Watts is also the director of Penn’s Computational Social Science Lab.
This article originally appeared in Annenberg School of Communication News.