False consensus in the echo chamber: Exposure to favorably biased social media news feeds leads to increased perception of public support for own opinions

Abstract

Studies suggest that users of online social networking sites can tend to preferably connect with like-minded others, leading to “Echo Chambers” in which attitudinally congruent information circulates. However, little is known about how exposure to artifacts of Echo Chambers, such as biased attitudinally congruent online news feeds, affects individuals’ perceptions and behavior. This study experimentally tested if exposure to attitudinally congruent online news feeds affects individuals' False Consensus Effect, that is, how strongly individuals perceive public opinions as favorably biased and in support of their own opinions. It was predicted that the extent of the False Consensus Effect is influenced by the level of agreement individuals encounter in online news feeds, with high agreement leading to a higher estimate of public support for their own opinions than low agreement. Two online experiments (n1 = 331 and n2 = 207) exposed participants to nine news feeds, each containing four messages. Two factors were manipulated: Agreement expressed in message texts (all but one [Exp.1] / all [Exp.2] messages were congruent or incongruent to participants' attitudes) and endorsement of congruent messages by other users (congruent messages displayed higher or lower numbers of “likes” than incongruent messages). Additionally, based on Elaboration Likelihood Theory, interest in a topic was considered as a moderating variable. Both studies confirmed that participants infer public support for their own attitudes from the degree of agreement they encounter in online messages, yet are skeptical of the validity of “likes”, especially if their interest in a topic is high.

Determinants of the backfire effect: Discrepancies between two groups with different political orientations

“…Taken together, the aforementioned factors induced the backfire effect of pan-green supporters. Highly politicised topics promote connections among individuals with similar political stances and encourage confrontation towards those with contradictory stances (Copeland and Boulianne 2020;Luzsa and Mayr 2021); this finding is consistent with those of the aforementioned studies.…”

“…When individuals observe high homogeneity in their environment in terms of their political ideology, they avoid being overly exposed to opposite narratives; this avoidance is conducive to deep political mobilisation (Boutyline and Wille 2016). Because of the large number of responses reflecting similar ideas, the members of polarised online communities may incorrectly believe that the general public agrees with their opinions and may lead them to take drastic actions, which sometimes can have severe consequences (Luzsa and Mayr 2021).…”

Social norms misperception among voters in the 2020 US presidential election

“…We predicted that at time 1, Republican supporters showed significantly higher approval than Democrat supporters for xenophobic statements and agreement estimation from “the wider American public”. To approach this, we first examined within effects (cf., Luzsa & Mayr, 2021; Sargent & Newman, 2021) by plotting and calculating the correlation between own approval and estimated agreement per party, followed by examining between‐group effects (cf., Watts & Larkin, 2010).…”

“…We predicted that at time 1, Republican supporters showed significantly higher approval than Democrat supporters for xenophobic statements and agreement estimation from "the wider American public". To approach this, we first examined within effects (cf., Luzsa & Mayr, 2021;Sargent & Newman, 2021) by plotting and calculating the correlation between own approval and estimated agreement per party, followed by examining between-group effects (cf., Watts & Larkin, 2010). The strength of association between approval for the necessity of harsh measures against immigrants and refugees and estimated agreement was stronger (and significant) for Republicans (r = .30, p = 0.03) than for Democrats (r = .07, p = 0.51), Figure 2A.…”

“…Sargent & Newman, 2021), some scholars see FC as present when own approval and perceived approval from others are associated (cf. Luzsa & Mayr, 2021), while others refer to the “traditional way,” comparing the strength of agreement perception between opposing stances (between‐group measurement; cf. Bauman & Geher, 2002; Ross, Greene & House, 1977; Watt & Larkin, 2010).…”