CUBIT

CUBIT’s Michaelmas 2020 seminar series, chaired by Dr Leor Zmigrod, explored political polarisation, a topic of academic discussion which has increasingly seeped into mainstream discussions of current affairs in recent years. So what is polarisation? At its core, political polarisation is a state of affairs where society is divided into two (or more) immiscible groups, each supporting political ideas the other regards as extreme, each holding hateful stereotypes of the other, and each embracing divergent versions of reality. Political groups in highly-polarised societies see each other, and each other’s ideas, as verboten, malign influences to be excluded and defeated, rather than constructive contributions to public debate. Unsurprisingly, polarisation has abysmal consequences for the societies it afflicts. Take Barack Obama’s summary of polarisation in contemporary America: “a fundamental contest between two opposing visions of what America is and what it should be; a crisis that has left the body politic divided, angry, and mistrustful, and has allowed for an ongoing breach of institutional norms, procedural safeguards, and the adherence to basic facts” (xv, A Promised Land).

Polarisation goes beyond people merely disagreeing about political issues, which is a natural consequence of liberal democracy. Rather, polarisation undermines democracy’s foundations. Ideally, democracy works something like a market, with politicians competing for citizens’ votes by proposing good ideas, ensuring that, over time, good ideas are rewarded and implemented, whereas bad ideas are rejected. But polarisation smothers this competition, as it leads people to vote for bad ideas simply because they are their side’s ideas, and, perhaps even more so, to reject good ideas just because they are the other’s.

So why is polarisation a matter for psychologists to study? When psychologists study polarisation, there is a tacit assumption that some features of human psychology make political disagreements even worse than they need otherwise be, even taking into account the very real and valid disagreements that political groups may have with one another about what is right, good, and moral. Psychologists do not contend that these differences exist only in people’s heads. Rather, we posit that the fundamentals of how people think about social groups amplify enmity and distrust. We suggest that how people think about political and moral issues can entrench division. Further, we hope that psychology can help with depolarisation – that if we can identify the aspects of human psychology that push us towards polarisation, we can use this knowledge to design interventions and redesign institutions to act as buffers against those forces.

So what can the abundance of research on polarisation tell us about how we polarise, and how we might depolarise? CUBIT’s though-provoking Michaelmas 2020 series offered some answers to these important questions, and raised yet more mysteries for researchers and policy-makers to wrestle with. The series consisted of four talks – from Dr Jeff Lees, Dr William Brady, Dr Andres Casas, and the trio of Ali Goldsworthy, Laura Osborne, and Alex Chesterfield – followed by a panel discussion involving Dr Lees, Dr Brady, and Dr Casas. Below, you can find the key questions and answers.

We assume our opponents are more hostile towards us than they really are
The series opened with a talk by Dr Jeff Lees, focusing on two recent projects he helped lead. His work concerns meta-perceptions – what people think other people are thinking. Lees presented evidence from a number of experiments where people were asked what people would think if a rival party attempted to spearhead some kind of non-partisan reform, like requiring judges to be elected. Importantly, people also indicated what they themselves would think in the same situation. 

The results were clear – people were less likely to oppose the reforms than people belonging both to their rival group and to their own group thought they would be. In other words, people often have an exaggerated view of how hostile their political opponents are towards them, and how hostile their political allies are towards their opponents. Such perceptions are the lifeblood of polarisation. The results were robust too, replicated across multiple countries and multiple scenarios. Lees was able to go even further, determining that the driver of this exaggerated perception of hostility is the belief that political groups compete with one another for influence, rather than cooperate with each other for the common good.

The implications for polarisation are clear—much partisan enmity springs from misperception. Simply by helping people to see each other’s views more clearly, we might reduce polarisation. Indeed, in an intervention experiment conducted by Lees’ team, where participants were told how opposed their opponents actually were to the reforms, they subsequently provided much more accurate meta-perceptions. Further, by reframing politics as a process by which people with different views cooperate and compromise for the common good, we might help depolarisation too. This may well require the implementation of new norms for political discourse, even institutional reform. It is perhaps telling that one of the countries with the lowest levels of polarisation in the modern world is The Netherlands, a country whose political system is so used to collaborative politics that from 1994-2002 and again from 2012-2017, the government was a ‘Purple’ coalition comprised of the leading left-wing and right-wing parties (VVD and PvdA). If the prospect of Labour and the Conservatives or Democrats and Republicans sharing the executive branch seems utterly unthinkable, perhaps the relationship between parties in the US and UK needs a rethink.

Social media can supercharge polarisation
The second talk of the series concerned the role of social media in polarisation, as was hosted by Rakoen Martens. Dr William Brady discussed an upcoming paper that explored how the design of social media platforms – namely Twitter – incentivises individuals to share polarising content online by encouraging the expression of moral outrage. There are two mechanisms. Firstly, a feedback mechanism, where individuals who express moral outrage in response to scandals receive more likes and shares than those who are more restrained, which incentivises them to carry on doing it. Secondly, a norms mechanism, where the online communities that people inhabit can develop a tendency to conspicuously express moral outrage, which encourages their members to do likewise. Why is moral outrage important for understanding polarisation? Prior research has shown that polarising information is more likely to spread if it contains moral content. More generally, the widespread expression of moral outrage entrenches polarised divisions.

Importantly, this research suggests that people don’t just express moral outrage online because they are morally outraged – they are incentivised to do so by the social approbation they receive, and encouraged to do so by the fact that others around them are doing it too. The very structure of social media acts like a reinforcement learning mechanism. Thus, while politically-engaged Twitter users no doubt frequently have valid reasons to express moral outrage, the design of the platform elicits such expressions, artificially inflating their presence in our online discourse, thereby polarising us further. While it is difficult to envision social media companies making the changes to their platform design required to neuter these effects, it is important for future real-world interventions to recognise that people are likely to engage in polarising behaviour because they get rewarded for doing so by their peers, and regard the behaviour as normal. Interventions will need to overcome the polarising inertia these twin mechanisms create.

A depolarised society is no utopia
The third talk of the series was held in the wake of the US election, with the hosts of the Changed My Mind podcast – Ali Goldsworthy, Alex Chesterfield, and Laura Osborne. The panel offered some important insights into the kind of society we should be aiming for as we attempt to depolarise. As Osborne put it, a depolarised society is not “one amorphous blob of shiny happy people” – division and disagreement still exist, but systems are organised so as to benefit from this diversity of opinion. In depolarised societies, people who disagree are able to communicate with one another effectively, and work together in good faith, with political leaders playing an important role of setting the example for how others should conduct themselves responsibly and constructively.

One crucial aspect of depolarised societies is that people are allowed to change their minds. The panel spoke of how they realised over their years in politics, academia, and the media, that once people become associated with a particular political stance in the US or the UK, they face very high costs to changing their mind, particularly in public. People who switch sides, or admit their prior beliefs were mistaken, are ridiculed much more often than they are applauded. Simply, if following the evidence is more costly than persevering with mistaken beliefs, we can only expect people to become increasingly unmoored from reality. In order to depolarise, the expectation that people should form and defend political opinions, almost for its own sake, should be replaced with the expectation that people’s political views will be amorphous, incomplete, and responsive to change.

How do we depolarise?
The penultimate talk of the series was hosted by Sakshi Ghai and featured Dr Andres Casas, a visiting scholar at the University of Pennsylvania, and Behaviour Change advisor to the Mayor of Bogota. Casas offered a glimpse of how we can successfully put depolarisation interventions into practice. Casas and colleagues set out to reduce polarisation between FARC rebels and wider society in his native Colombia. This was a formidable challenge, seeking to overcome a division fuelled by 60 years of brutal civil war. The group’s prior work had even revealed that no group of people is viewed in a more dehumanised way than FARC militants are by ordinary Colombians.

But the researchers’ approach was strikingly simple: first, they set out to discover what thoughts and feelings led ordinary people to oppose making peace with FARC, then they set out to ‘unfreeze’ those thoughts and feelings, by designing an intervention to change them directly. They discovered that the perception that FARC were unwilling or unable to integrate into Colombian society was crucial to underpinning opposition to FARC reintegration. To change people’s minds about this perception, they travelled to a town (Dabeiba) situated next to a camp for former FARC militants, where integration had been encouraged (under military supervision), and the communities now shared amenities like a school and a football team, to interview people on both sides about their perceptions of the other, then showed the interview to ordinary Colombians. The interviewees spoke of how well the two communities were integrated, and watching the video significantly increased people’s support for FARC reintegration, primarily by increasing their perception that FARC were willing to integrate.

There are three things we can learn from this successful intervention. Firstly, it underscores the notion that often polarisation is driven by mistaken meta-perceptions – beliefs about what the ‘other side’ are thinking. Left unchallenged, meta-perceptions will often be excessively negative. Second, it reinforces the principle that simply showing people the truth about what other people believe can increase the accuracy of these meta-perceptions, and thereby reduce polarisation. These first two points are consistent with the work by Dr Lees. Third, this intervention demonstrates effective design principles for future depolarisation efforts – use primary research and theory to identify the thoughts and feelings which drive polarisation in a particular population, then implement an intervention to challenge those thoughts and feelings directly.

Connecting the dots
The last event of term brought Dr Lees, Dr Brady, and Dr Casas back together for an integrative discussion of their work. The discussion highlighted the elusive nature of polarisation, and pointed out some further points of agreement, as well as questions for future research.

So, to start with the points of agreement – it is clear that social identity is a key driver of polarisation. As the Changed Your Mind panel had pointed out, it is when individuals feel that a particular political identity is connected to their sense of self that they become most in danger of polarisation, as they simply do not want to be seen to change their minds. Therefore, the panel agreed, reducing the saliency of group identities in political discussions can aid depolarisation efforts.

Further, narratives are key. Narratives are simplified stories people tell each other about what is going on in the world and why. Often, our narratives about politics lead to polarisation. It is hard to explain other people’s behaviour without referencing their intentions, and as Dr Lees’ and Dr Casas’ discussions of meta-perceptions showed, we are often excessively negative in our judgment of rival political groups’ intentions. Therefore our narratives about them become inaccurate and polarising. This tendency becomes self-reinforcing too, as Dr Casas pointed out. As we separate ourselves from the groups we dislike and distrust, we have fewer opportunities to learn about each other accurately, making our narratives increasingly misguided. A kind of cognitive path dependence kicks in too – as we share ideas among our group, but not those in other groups, what one group holds to be true and right can become very distinct from what the other group holds to be true and right, creating new sources of real disagreement, pushing the groups further apart, even if the initial divide between the two groups is only subtle. But while narratives can pull us apart, Dr Casas noted that narratives can bring us together too – like the simple video telling the story of how FARC and ordinary Colombians had come together in Dabeiba.

What about questions for future research to approach? One question was whether polarisation is inevitable. Perhaps with vigilance over our meta-perceptions, our norms, our narratives, the role of identity, and the design of our social media, it will be possible to contain polarisation at a manageable level through the appropriate use of interventions and design of political institutions. But there certainly are, and always will be, strong psychological forces that predispose us towards polarisation.

But the big question is this – what really is polarisation? As the panel pointed out, it is primarily media interest in the phenomenon of polarisation that drives academic research, but in truth the body of literature on polarisation is still in need of formal organisation. We know about many of the facets of polarisation, we have many measures of it, and we have many theories about how it emerges. But there remains the sense that when we talk about polarisation, we might all be talking about slightly different things (it has recently been suggested that this issue – of applying intuitive but imprecise labels to heterogeneous phenomena – is in fact a much broader problem within psychological research). Is there a unifying essence of polarisation, something discrete and measurable to which its dangers can be attributed beyond anything else? In other words, what are the minimal criteria that separate a polarised society from a non-polarised one? Is it the extremity of the difference between the positions adopted by the two major political blocs in a society? Is it the mutual loathing between the groups? Is it the unwillingness of the groups to compromise with each other? Or is it just a natural and temporary state of affairs that arises when societies find themselves divided over important issues? While this CUBIT talk series showed that answering these questions need not preclude us from making progress tackling polarisation, they are questions that remain to be answered. These are mysteries for future research to resolve.

By David Young (@DavidJ_Young), with special thanks to Lee de-Wit (@LeedeWit), Leor Zmigrod (@LeorZmigrod), and Rakoen Maertens (@RakoenMaertens) for the helpful suggestions and assistance.