Examining the Role of Source Information in Citizen-to-Citizen Political Appeals

• In Event: Cues vs. Content: How People Reason about Politics

Abstract

Political persuasion is a more citizen-driven enterprise than it used to be. Twenty years ago, political messages were carried to the masses via television, radio, and newspapers—and not much else. All of these mediums center messages from a fairly small number of prominent elites: politicians, pundits, spokespersons, and the like. In contrast, today’s media environment offers myriad opportunities for regular citizens to engage directly with each other: they comment on Facebook posts, dispatch tweets on X, and improvise commentary about the news on TikTok. It is not uncommon for messages sent via such routes to be seen by hundreds or thousands of others.



Our paper is part of a broader project seeking to update theories developed to understand elite-to-citizen communication for the world of citizen-to-citizen communication. The current paper examines the role played by partisan cues. For elite-to-citizen communication, party affiliations are often obvious and easily accessible. (A politician’s party affiliation is often common knowledge. Even when not, television networks routinely include a person’s party affiliation right on the chyron.) For citizen-to-citizen communication, the same affiliations are not always so obvious. Some social media users boldly display their political affinities in their profiles, but others do not, leaving it for message recipients to infer political leanings, if they can.



To understand when citizen-to-citizen persuasion efforts might be facilitated or thwarted by partisan considerations, we develop expectations for how persuasion should depend on 1) the presence or absence of a partisan cue and 2) the topic under discussion. A key distinction in our setup is between consolidated issues—ones where partisan divisions are salient and widely known—and non-consolidated issues—ones where issue positions do not neatly map onto party affiliations. We hypothesized that party cues would color perceptions of non-consolidated issues more than for consolidated issues.



We tested this hypothesis via a pre-registered randomized experiment wherein hundreds of participants wrote sincere political appeals attempting to persuade others on four pre-determined policy issues: two of which were consolidated and two non-consolidated. We then presented these messages to political opponents, randomizing whether they were presented with or without information about the party identity of the author. A virtue of this naturalistic design is that it allows us to disentangle message effects—the recipients’ positive or negative reactions to what a person wrote—from the cue itself. That is, we can distinguish the extent to which messages are penalized because of their content, versus the mere fact of who wrote them.



The first iteration of this experimental paradigm challenged our expectations, since we found comparable persuasion on both consolidated and non-consolidated issues. Of particular interest, we uncovered one issue (out of four we examined) where, against our expectations, providing out-party cues improved recipients’ perceptions of a message. These results suggest that perceptions of citizens’ messages are more complex and context-dependent than we anticipated.



Our next experiment, slated to be fielded in February, tests explanations for Study 1 patterns that ran counter to our expectations. In particular, we examine the role played by partisans’ often exaggerated caricatures of what issue stances are typical of out-party members. To test a new hypothesis that out-party appeals are more successful when they defy a person’s preconceptions, we measure respondents’ perceptions of the typical out-party member’s issue views and then manipulate whether respondents consider arguments that are typical or atypical.



According to work on cognitive fluency, we expect that respondents will have an easier time processing persuasive arguments that are harmonious with previous beliefs about the party’s issue stance. We also expect consolidated issues to exhibit established and broadly identifiable partisan cleavages. The usefulness of seeing partisan cues should thus be dependent on whether respondents would be able to accurately identify the partisanship of the sender in the absence of such cues. These partisan cues should be most jarring to respondents with stronger partisan ties and those who have higher levels of political knowledge, both of whom may have assimilated a caricature of the out-party members’ issue stances. We expect that these results will help clarify and further our understanding of the persuasive effects that ran counter to our expectations in the first study.

Authors

• Connor Choate, University of North Carolina, Chapel Hill

• Allen Wilson

• Megan Rickman Blackwood, University of North Carolina, Chapel Hill

• Timothy Ryan, UNC Chapel Hill