When two people with opposing political beliefs are shown the same piece of information, they often have radically different responses. Drawing on new research, Jens Koed Madsen, David Young and Lee de Wit show that disagreements over the independence of information sources can play a significant role in this type of polarisation.
Research shows that people with opposing political beliefs can be exposed to the same evidence from the same source and update their beliefs in opposing directions. At first glance, this kind of “belief polarisation” might seem illogical. Indeed, it has often been ascribed to irrational reasoning. But there are reasons to think that it might be more rational than it first appears.
As psychologists, we are interested in how belief polarisation might come about – both because we want to understand what causes it and so that we can advise governments and NGOs on how we might reduce it and foster more constructive democratic debates. For this reason, we are interested in whether this polarisation can occur from purely rational belief revision processes.
That is certainly not to say that all political behaviour is rational, but it is important for public policy to try to identify elements of polarisation that might reasonably arise from logical cognitive processes in complex media environments. Indeed, we have recently argued that behavioural public policy should start from an assumption that people are reasonable and rule this out before we can claim essential differences or irrational biases. This is also known as the Principle of Charity.
Dependencies and polarisation
In a recent study, we examine whether people’s beliefs about the sources of information can lead to issue polarisation. Specifically, we assess their perception of source dependencies. A lot of work has focused on whether people trust particular sources more than others, but what we are interested in is the perceptions people have of the relationships between sources of information.
Sources can be perceived to be independent of each other. If two sources that agree are completely independent of each other, they have reached the same conclusion without talking to each other, looking at the same evidence, or seeing each other’s reports. Dependent sources in contrast may share evidence or a way to interpret data, they may have discussed a case before making a statement, or they may both wish the results turn out some specific way.
Across many domains of psychology, we see that people are sensitive to whether sources are perceived to be independent. Sources perceived to be more dependent should carry less weight than reports from sources that are independent. This has two implications for polarisation that we test in our study.
First, we examine whether manipulating people’s perception of dependence can generate polarisation. Second, we study whether people on opposite sides of the political spectrum actually believe that the other “side” is more dependent than their own side, which would allow them to explain why others look at the same evidence and reach opposite conclusions.
Dependency matters
To test the first idea, we ran an experimental study showing that people do polarise on issues when they differ in their dependency beliefs. Participants considered whether a politician was guilty of embezzlement under scenarios in which different witnesses stated that he was guilty or innocent.
We found participants were more likely to believe the politician was guilty when they were told that witnesses who claimed the politician was guilty had similar backgrounds, read the same newspapers and adhered to the same ideological position – and vice versa. The experiment showed us that simply manipulating people’s perceptions of dependency was sufficient to create polarisation.
To test the second insight, we tested whether people with different political convictions ascribed higher dependencies to their political out-group than their in-group. In a UK survey, we asked people to rate how dependent they thought people who voted Labour and Conservative were (as well as their own political preference so that we could group participants).
We developed a 9-item survey that asked about aspects like information consumption, ideological position, goals, life experiences, opinion transmission and more. For example, on information dependency, participants rated Labour and Conservative voters from 1 (“they all get their information about politics from the same source”) to 7 (“they all get their information about politics from different sources”).
We found a strong effect of dependency across all 9 items, where participants on both sides of the political divide thought their in-group was more independent and the out-group was more dependent (Figure 1). This also held when we replicated the survey in the United States for Democrats and Republicans (Figure 2).
Figure 1: Dependency ratings for Labour and Conservative supporters
Note: For more information, see the authors’ accompanying paper in Cognition.
Figure 2: Dependency ratings for Democrat and Republican supporters
Note: For more information, see the authors’ accompanying paper in Cognition.
The findings support the idea that people from both political sides believe that the “other” side consumes information from a structurally more dependent ecosystem. This makes it easier for people to explain why other citizens disagree with them on issues.
If the “other” side truly did get their information from a more dependent system, it would be subject to more error and therefore their positions should be taken less seriously. The problem is that both sides seem to strongly believe this about each other, which is less helpful.
Tackling polarisation
Our findings have practical implications. First, they suggest we should be more cautious in simply assuming that that people from “the other side” are just irrational and that we need to understand their perception of the information ecosystem. Our findings suggest that there is space for discussion and dialogue, which is a positive democratic view.
Second, the findings point to the types of conversations that would be needed to de-polarise people. If people truly disagree on the reliability and dependencies of sources, they should not be persuaded by additional evidence from those sources. Instead, the conversations should be about why we trust the sources we do and how we see their relationships.
This has practical implications for broader beliefs like climate change denial. If deniers earnestly believe that climate scientists are unreliable and that they are all working together to push a particular agenda, additional reports from scientists should not move their beliefs. Instead, scientists need to show how the scientific process encourages independent studies.
Finally, the model we propose in the paper suggests a polarisation cycle, which needs to be understood. That is, if a person’s prior beliefs align with the explanation given by one side of the political spectrum, they can explain that others disagree by inferring a higher degree of dependency.
However, if the person now believes the other side is more dependent, they should update less when they see reports from that side, thereby privileging reports from their “own” side on new issues. This would lead to a back-and-forth where citizens with different starting points may polarise on different issues because they update their beliefs about the ecosystem rather than just on the hypothesis in question.
Of course, the causes of polarisation are more complex and may involve different systemic and individual factors like social media algorithms, media consumption, biases, social networks, economic circumstances and more. But we think it is important to avoid neglecting the cognitive beliefs of individual citizens and how their perceptions of the information environment might shape the direction in which they update their beliefs on political issues.
Our work complements recent findings that people are less likely to listen to experts if they do not think they are independent, and suggests that an important feature of healthy democratic media ecosystems is that there are independent sources of information out there which are not just sticking to “the party line”.
For more information, see the authors’ accompanying paper in Cognition. The research was funded by the Templeton World Charity Foundation (ID: 31453) with the help of Nick Barron from MHP for data collection.
Note: This article gives the views of the authors, not the position of EUROPP – European Politics and Policy or the London School of Economics. Featured image credit: Lightspring / Shutterstock.com