debate twitter video.
Stance polarity in political debates: A diachronic perspective of network homophily and conversations on Twitter.
.png)
Introduction
Social media have changed the information consumption and diffusion behavior, as in [1], [2], gaining a crucial role in the public debate about socio-political issues in both directions, as in [3]: from institutions and politicians to citizens (top-down), and conversely (bottom-up). Indeed, some political leaders make an extensive use of platforms like Twitter or Facebook to communicate with citizens, e.g., in [4], [5], [6], that, on the other hand, join in online discussions supporting or criticizing their political opinions. In this framework, social media provide a powerful tool to test the public opinion mood and investigate how individuals are exposed to diverse viewpoints. Developing automated systems for a deep analysis of users’ generated contents and interactions is becoming increasingly relevant, and recent works focused on detecting users’ opinion towards a particular target, e.g., in [7], [8], [9], [10].
Recent studies suggest that web users tend to polarize their opinion and form partisan political communities, e.g., in [11], [12], following the homophily principle in [13], [14], according to which like-minded people are more likely to connect to each other. Despite the scientific debate is still open about the role of the architecture of these platforms in the formation of social groups, as in [15], [16], many scientists suggest that the presence of echo chambers (i.e., when users are exposed only to information from like-minded ones) and filter bubbles (i.e., when content is selected by algorithms according to the user’s previous behaviors) reinforce people’s pre-existing beliefs, filtering and censoring divergent viewpoints, as in [17], [18]. Moreover, Sunstein in [19] suggests that two persons, who only slightly disagree with each other, are likely to be even more opposed, after they have talked to each other, while democracies should be based on a conciliation among viewpoints.
Abstract
In the last decade, social media gained a very significant role in public debates, and despite the many intrinsic difficulties of analyzing data streaming from on-line platforms that are poisoned by bots, trolls, and low-quality information, it is undeniable that such data can still be used to test the public opinion and overall mood and to investigate how individuals communicate with each other.
With the aim of analyzing the debate in Twitter on the 2016 referendum on the reform of the Italian Constitution, we created an Italian annotated corpus for stance detection for automatically estimating the stance of a relevant number of users.
We take into account a diachronic perspective to shed lights on users’ opinion dynamics.
Furthermore, different types of social network communities, based on friendships, retweets, quotes, and replies were investigated, in order to analyze the communication among users with similar and divergent viewpoints. We observe particular aspects of users’ behavior.
First, our analysis suggests that users tend to be less explicit in expressing their stances after the outcome of the vote; simultaneously, users who exhibit a high number of cross-stance relations tend to become less polarized or to adopt a more neutral style in the following phase of the debate. Second, despite social media networks are generally aggregated in homogeneous communities, we highlight that the structure of the network can strongly change when different types of social relations are considered.
In particular, networks defined by means of reply-to messages exhibit inverse homophily by stance, and users use more often replies for expressing diverging opinions, instead of other forms of communication. Interestingly, we also observe that the political polarization increases forthcoming the election and decreases after the election day.
In our previous studies, we analyzed two political debates on Twitter focusing on two events that have been considered as symptoms of a new nationalist populism in [20]: Trump’s election and Brexit. In our first analysis, we only focused on linguistic aspects with the aim to determine from the text of a tweet whether the author is in favor of the given target, i.e., Hillary Clinton, Donald Trump in [21].
Analyzing the political debate around the so called Brexit referendum in [22], we inspected the debate at user level by aggregating, in a diachronic perspective, the tweets posted by the same user in a single day.
Information of existing follower/followee relations were analyzed and used to create network-based features that improved stance detection performance.
In this study we examine the political debate on Twitter about the Italian constitutional referendum held on December 4th, 2016 in Italy, adopting the machine learning model we obtained previously in the same scenario in [23]. Some political analysts tried to explain the result of the Italian constitutional referendum, similar to Brexit and Trump’s election results, as a reaction of a sense of disbelief against the elites. The international pressure on this case arose when many influential actors expressed their support for (i.e., JP Morgan, Fitch, the Financial Times, the Wall Street Journal, and the 44th USA President Barack Obama etc.) or criticism of (i.e. the Economist) the constitutional reform as a way to stop the spreading of populism in western democracies. Referring to a national report about the Italian social situation in 2016 in [24], [25], a not negligible portion of Italian people use Facebook and get news from them:
of interviewed people declare to use Facebook as their main source of information, as opposed to the
and
of Italians that use TV and Radio news outlets. Although Twitter is not the dominant source of information for Italians, there is still a
of citizens that declare to use Twitter (
among the young population) for such a purpose: apparently, studying opinion dynamics in data streaming out from such a platform is likely to return a signal of what is going on in the more general public debate. We are aware that a percentage of approximately
of citizens is not representative of the whole population’s stances; in fact, we want to remark that our study is limited to how Italian users relate to Twitter when they expose their stances towards a given topic, and that these results are not necessarily generalizable to other communication networks commonly used.
Our aim is to show that such techniques can at least complement traditional political polls, that are more accurate in terms of the selection of socio-demographic features of the sample, but that relies on a much smaller set of citizens that are directly interviewed by poll agencies. For example, it has been reported (source: https://www.ilpost.
it/2018/01/21/sondaggi-elezioni-2018/) that during 2018 political elections in Italy, a number of telephone interviews varying from 259 to 850 has been placed to collect statistical data.