Metacognitive effects of attitudinal (in)congruence on social media: relating processing fluency, subjective knowledge, and political participation

Abstract

Introduction: Encountering political disagreements in our daily lives can discourage us from participating in democratic processes. To date, research has mainly focused on social motives or attitudinal mechanisms to explain this phenomenon. In the present study, we adopt a different approach and highlight metacognitive effects of attitudinal (in)congruence on processing fluency (i.e., perceived ease of processing) and subjective knowledge as well as their relationship with behavioral outcomes such as the intention to politically participate.

Methods: In a pre-registered online experiment (N = 1,258), participants saw a political social media post with six opinionated user-generated comments. These comments either all matched (congruent condition) or contradicted (incongruent condition) participants' personal opinions. Processing fluency, issue specific subjective knowledge, and intention to politically participate were then measured using established self-report scales.

Results: In line with our hypotheses, the congruent stimuli evoked a higher feeling of processing fluency than the incongruent ones (d = 0.21). Furthermore, participants in the congruent condition indicated a higher intention to politically participate (d = 0.23) and rated their own knowledge on the topic as higher (d = 0.22) than participants in the incongruent condition-even though the factual knowledge gain should be equal in both conditions. Finally, we observed positive relationships between processing fluency and subjective issue knowledge (β = 0.11) as well as between subjective issue knowledge and issue-specific political participation (β = 0.43).

Discussion: Our findings highlight the importance of considering metacognitive constructs such as subjective knowledge to explain political behaviors and suggest that attitudinal congruence influences the way we perceive our own knowledge and information processing.

Keywords: attitudinal congruence; cross-cutting exposure; metacognition; political participation; processing fluency; subjective knowledge.

Figure 1

Overview of the overall theoretical model.

Figure 2

Original stimulus material with English translations in gray areas. Participants saw either opposing or supporting comments.

Figure 3

Distributions of prior opinions. Participants who chose answering option 4 (indicating no clear opinion) were further asked what their opinion would be if they had to decide (not included in this figure).

Figure 4

Violin plots for group differences in (A) political participation and (B) subjective knowledge. All values are mean-centered within their respective topic (original scale: 1–7). The black dots indicate the mean with its 95% confidence interval as error bars. The blue shaded area displays the kernel density plot of the data.

Figure 5

Mean differences between main topics and secondary topics regarding (A) political participation and (B) subjective knowledge. Main topic indicates that the participants saw a post with opinionated comments to this topic; secondary topic indicates that the participant saw a post without any comments to this topic (classification of congruent vs. incongruent is based solely on the main topic stimulus of each participant). Error bars indicate the standard error of the mean; scale: 1–7.

Figure 6

Results of regression analyses. (A) Linear regression of subjective knowledge on political participation. All values are mean-centered within their respective topic (original scale: 1–7). The shaded area indicates the 95% confidence interval. (B) Logistic regression of subjective knowledge (mean-centered) on whether or not the participants wish to be invited to an online discussion event. The regression line indicates the predicted probability of choosing yes (i.e., wishing to be invited). The vertical gray lines indicate the area of observed subjective knowledge values. The shaded area indicates the 95% confidence interval.

Figure 7

Results of mediation analysis. RMSEA = 0.10 [0.05; 0.15]; CFI = 0.96; SRMR = 0.03; ** denotes p < 0.001, * denotes p < 0.05.