Social and political polarization is one of the challenges facing contemporary societies. The Heurística association and the Tecnopolítica unit of the the UOC IN3 Communication Networks & Social Change (CNSC) research group organized the round table “Do algorithms dream of polarized sheep? Social and political polarization and social media” as part of the 2nd DataPolitik conference, that took place on May 18 and 19 on politics and communication in the big-data era. The sessions were an invitation to transdisciplinary dialogue on the transformations of communication and the grammars of social interaction, the result of the popularization of digital platforms, convened to discuss these issues and promote deliberation on the internet.
Borge began by contextualizing the situation in Spain, highlighting that surveys indicate that both political and affective polarization have increased in recent years and are above the average for Europe, but below the levels seen in the United States. She explained how the aggregate data show peaks of political polarization in 2008, with the beginning of the economic crisis; in 2015-2016, with the end of two-party system and the entry of Podemos and Ciudadanos into the Spanish parliament; and in 2019 with the territorial crisis of Catalonia and the rise of VOX. Despite this upward trend, she noted that “a certain degree of conflict and political division is positive for the party system, the public sphere, and for democracy in general”.
The danger of affective polarization for democracy
She highlighted that the greatest danger for democratic societies lies in the increase in affective polarization. That is, when political differences become “hatred of those people who are not from the same party or do not share the same ideological perspectives, that’s a situation that can trigger intolerant and violent attitudes”.
Muñoz emphasized the impact of affective polarization and described an experiment carried out by his research group in which they analysed the extent to which political polarization can erode the ability of citizens to censure anti-democratic behaviour. The results showed that ideological proximity did not increase tolerance of anti-democratic behaviour; in fact, people tended to be more disapproving of those whose ideological outlook was similar to their own. Indeed, he explained that “in terms of affective polarization, there was greater tolerance of this type of anti-democratic behaviour when carried out by people from a different ideological profile to their own”.
The opacity of social media algorithms
The attendees also pointed to social media, especially Twitter and Facebook, as one of the important factors in polarization. “It’s like a vicious circle, people want to relate to like-minded people, on top of which social media, by their nature, reinforce this trend with algorithms that show us opinions similar to our own,” explained Borge. Faced with this reality, Castillo expressed his concern about the “opacity” of these algorithms and the difficulties in exercising “democratic control over these platforms”.
However, Borge also pointed out that polarization on these platforms may be “overstated”. “First, not everyone is on Twitter or Facebook or other social media all the time, as many people are just occasional users who dip in for entertainment or to find information. And, also, the vast majority of people have moderate positions, do not avoid content that does not match their opinion and do not usually block people with differing opinions. In fact, some studies’ surveys indicate that 67% interact with politicians or parties with an ideological background different to their own,” she said.
Likewise, Muñoz pointed out that perhaps the vision of social networks as “echo chambers” is not entirely correct, as, in his opinion, “we are more exposed to disagreement on social media than in everyday life,” where our friends usually share our own views.
The negative effects of exposure to disagreement on social media
Muñoz also noted that perhaps we should modify the idea, supported by classical studies on the subject, that exposure to disagreement always benefits deliberative processes. For him, this may be the case in the analogue world, but “perhaps this exposure does not have the same effect on social media,” especially in an environment of affective polarization in which the biases with which we process information are accentuated. “The kind of exposure to disagreement that occurs on social media does not trigger the positive effects of empathy, of putting yourself in the other’s shoes, of reaching agreements, etc., that take place when we discuss matters with others in person. The way in which you process information on social media is highly conditioned by your political preferences, activating biases such as the reduction of cognitive dissonance and confirmation bias, which diminish the positive effects of exposure to disagreement on the deliberation process.”
Education and collaboration
The speakers also outlined some of the measures needed to address this issue in the current context, both from the point of view of citizens and research. Borge stressed the need for digital literacy and collaboration among citizens. “We must continually educate ourselves in the digital field and collaborate with each other. Social media create isolated worlds, and we must help each other, especially between generations, so that people who do not have access or adequate knowledge do not remain isolated.”
Muñoz emphasized the need for collaboration, but in the field of research, connecting “the different areas of research and the different methodologies to form a more complete vision” of the reality of polarization and social networks.
Finally, Castillo stressed the importance of improving technological designs and transparency of algorithms, as this is a “key aspect for democracy and we are leaving it in the hands of computer scientists with all their biases”.