The preference for belief, issue polarization, and echo chambers

Abstract

Some common explanations of issue polarization and echo chambers rely on social or cognitive mechanisms of exclusion. Accordingly, suggested interventions like "be more open-minded" target these mechanisms: avoid epistemic bubbles and don't discount contrary information. Contrary to such explanations, we show how a much weaker mechanism-the preference for belief-can produce issue polarization in epistemic communities with little to no mechanisms of exclusion. We present a network model (with an empirically-validated structure) that demonstrates how a dynamic interaction between the preference for belief and common structures of epistemic communities can turn very small unequal distributions of initial beliefs into full-blown polarization. This points to a different class of explanations, one that emphasizes the importance of the initial spread of information. We also show how our model complements extant explanations by including a version of biased assimilation and motivated reasoning-cognitive mechanisms of exclusion. We find that mechanisms of exclusion can exacerbate issue polarization, but may not be the ultimate root of it. Hence, the recommended interventions suggested by extant literature is expected to be limited and the problem of issue polarization to be even more intractable.

Keywords: Echo chambers; Epistemic bubbles; Homophily; Issue polarization; Mechanisms of exclusion; Motivated reasoning; Preference for belief; Preference for certainty.

Fig. 1

Motivated bias can both produce and amplify existing polarization. Notice, however, that some polarization can happen even when there is no motivated bias, as shown in plots in the right column under Arranged Start. The top row shows simulations done on random graph networks, the bottom row on small world networks. Actual simulated levels of motivated bias were 0, 0.25, 0.5, and 0.75. Each parameter setting was simulated 200 times

Fig. 2

Results from simulations where two “Yes” beliefs and two “No” beliefs were initially distributed in different ways. Upper case letters indicate a token belief-yes, lower case letters a belief-no, and the type of letter indicates ideology type (M = moderate with belief-yes, o = (orange) right ideology with belief-no, P = (purple) left ideology with belief-yes, etc). “Noise” on the x-axis is for illustrative purposes only to avoid dots completely covering one another. Each parameter setting was simulated 1,000 times

Fig. 3

Results from simulations where three “Yes” beliefs and one “No” belief were initially distributed in different ways. (Notation as above in Fig. 2.)

Fig. 4

Incongruent starts. Some agents are given beliefs that are incongruent with their ideology and we then explore how much motivated bias is needed to result in belief spread that is largely congruent with ideology. For strong ideologies (Very Left and Very Right) this happens at about 62.5%, while for the weak ideologies (Left and Right) this happens at about 75% (reminder: that is the level that the strong ideologies reject, while weak ideologies are half that). Actual levels of motivated bias ranged from 0 to 1 in 0.1 increments. Each parameter setting was simulated 1,000 times

Fig. 5

Results from our robustness check simulations where two “Yes” beliefs and two “No” beliefs were initially distributed in different ways. All distributions are nearly identical with the exception of ppOO with Bias, where here we no longer see polarization as before

Fig. 6

Results from our robustness check simulations where three “Yes” beliefs and one “No” belief were initially distributed in different ways. Notice in particular the y-axis: now agents are approaching consensus

Fig. 7

Incongruent starts. Some agents are given beliefs that are incongruent with their ideology and we then explore how much motivated bias is needed to result in belief spread that is largely congruent with ideology—here using our local majority updating version