How Does Belief Polarization Work?

by Scott F. Aikin and Robert B. Talisse

We have noted previously that there are two different phenomena called “polarization.” The first, political polarization, refers to the ideological distance between opposing political parties. When it’s rampant, political rivals share no common ground, and thus cannot find a basis for cooperation. Political polarization certainly poses a problem for democracy. Yet belief polarization is perhaps even more troubling. It is the phenomenon by which interactions among likeminded people result in each adopting more radical versions of their views. In a slogan, interactions with likeminded others transform us into more extreme versions of ourselves.

Part of what makes belief polarization so disconcerting is its ubiquity. It has been extensively studied for more than 50 years, and found to be operative within groups of all kinds, formal and informal. Furthermore, belief polarization does not discriminate between different kinds of belief. Likeminded groups polarize regardless of whether they are discussing banal matters of fact, matters of personal taste, or questions about value. What’s more, the phenomenon operates regardless of the explicit point of the group’s discussion. Likeminded groups polarize when they are trying to decide an action that the group will take; and they polarize also when there is no specific decision to be reached. Finally, the phenomenon is prevalent regardless of group members’ nationality, race, gender, religion, economic status, and level of education.

Our widespread susceptibility to belief polarization raises the question of how it works. Two views immediately suggest themselves, the informational account and the comparison account.

On the first, discussion with likeminded others exposes us to a high concentration of affirming reasons and ideas. Moreover, in such contexts, there is typically a scarcity of countervailing or disconfirming considerations. Consequently, group members absorb the new information, and revise their own view in light of it. As the new information confirms their antecedent view, they become more extreme advocates.

Although the informational account surely captures part of what drives belief polarization, it cannot be the entire story. For one thing, belief polarization has been found to occur in groups even when new and novel information is not presented. In fact, it has been found to occur even in contexts where group interactions involve no exchange of information at all.

This suggests an alternative, the comparison view, which holds that belief polarization results from in-group comparisons. Group members care about how they are perceived by the other members. In the course of discussion, they get a better feel for the general tendencies within the group, and, wanting to appear to others neither as a half-hearted outlier nor as an over-the-top fanatic, they update their opinions so that their view lies notably above what they perceive to be the mean, but beneath what they regard as unacceptably hardline. Now, given that group members are engaging simultaneously in this kind of recalibration, the tendency to escalating extremity is to be expected.

Although more promising than the strictly informational view, the comparisons account is still lacking. Just as belief polarization can occur in the absence of the exchange of information, it can be induced in the absence of in-group comparisons, too. Indeed, the phenomenon can be activated even in the absence of anything that would count as interaction among the members of the likeminded group. It is not real time comparisons that drive the phenomenon so much as the subject’s own internal estimations of the dominant tendencies within his or her identity group. So neither information-exchange nor in-group comparison is strictly necessary for the effect; rather, the relevant shifts occur simply in light of group-affiliated corroboration of one’s views. That is, belief polarization can occur simply when an individual is caused to feel that a group with which she identifies widely shares a view that she espouses. She need not hear any reasons in favor of the view, nor need she be in the presence of other members of the group with whom she can compare herself. Instead, the realization that one’s belief is popular among one’s identity group suffices for belief polarization.

Thus a third account, the corroboration view, holds that shifts towards extremity can occur simply as a result of in-group corroboration. Corroboration from our peers makes us feel good about our shared beliefs, and this makes us feel affirmed in our social identity. In turn, when we feel affirmed in this way, we shift towards extremity.

So it turns out that although belief polarization predictably occurs in discussion among likeminded people, this is not necessary for the effect. Indeed, corroboration can come by way of highly indirect channels. For example, presenting a subject who identifies as liberal with a chart showing that liberals widely oppose genetically modified food can prompt belief polarization. And exposure to a poll showing that conservatives overwhelmingly favor a particular military action can produce an extremity shift in the belief content of a conservative already favorably disposed to that action.

An intriguing implication follows. The social environment itself can trigger extremity shifts. These prompts need not be verbal, explicit, or literal; they can be merely implicit signals to group members that some belief is prevalent among them – hats, pins, campaign signs, logos, and gestures are all potential initiators of belief polarization. Further, as corroboration is really matter of numbers, those with the power to present the appearance of widespread acceptance among a particular social group of some idea thereby have the power to induce extremity shifts among those who identify with that group.

It is sometimes claimed that the proper response to belief polarization is to diversify our sources of information. This is of course a good idea in any case. However, the corroboration view of belief polarization suggests that this measure is insufficient. Irrespective of the choices we make to expand our informational exposure, the physical and social surroundings that we inhabit in our day-to-day life can transform us into more extreme versions of ourselves.