Examining the EU citizens’ sentiments about the EU security and defence policy through Twitter: The case of the Ukraine crisis

Ioannis Galariotis (SPS 2015-2016) (*)

One of the fundamental aims of political scientists is to develop robust methodological frameworks in order to explain, understand and critically assess political phenomena. The discipline of establishing causality among dependent and control variables has been deemed for years the key theme of dispute but also of theoretical/methodological elaboration in the political science literature. Recent developments in the field of computational social science (CSS) approaches come to substantially enlarge the existing methodological techniques for the establishment of causality and, consequently, for the better understanding of political phenomena.

Despite a wider critique that computer science approaches receive from the camp of social scientists regarding their applicability to explain social phenomena, it is fair to make clear that CSS approaches do not comprise competitive or alternative methodologies compared to the existing ones being usually used in political science, but they have to be perceived as complementary in the sense that they aid researchers in the social science fields to improve and reinforce their methodological arsenal. However, problems of communication and understanding between the researchers of the two disciplines, political science and computer science, have not left them to realize that they have common knowledge to share for the formation of robust methodological frameworks and the improvement of theoretical building.

My first research article as a Max Weber Fellow (2015-2016) at the Department of Political and Social Sciences deals with the issue of how the citizens of the EU member-states conceive the EU as a global security actor and what their perceptions and sentiments are about the EU security and defence policy. Do the citizens of the EU member-states feel convinced (and/or angry) about how the EU behaves in world affairs? And, to what extent can the EU be conceived as a global security actor by its own citizens?

Eurobarometer surveys comprise one of the most reliable sources we have in our hands to estimate the perceptions of the EU citizens with regard to the issue of the EU security and defence policy. Whatever the estimated methods are being used by the Commission to track the public opinion, there are powerful computational sentiment techniques that have been tested to various disciplines (such as within the marketing and business fields) in order to trace and estimate the sentiments and perceptions of the public opinion with regard to a particular issue. These sentiment analysis methods are part of a wider methodological bundle of natural language processing techniques and CSS approaches.

To estimate and examine the perceptions and sentiments of the citizens of the EU member-states, this paper will use data from the Twitter regarding the role of the EU in the recent Ukraine crisis. Millions of tweets in the English language will be explored and analyzed from February to April 2014, when Russian troops invaded in the Ukrainian autonomous region of Crimea resulted in the annexation of Crimea by Russia on the 18th of March 2014. Although data stemming from Twitter may be characterized as biased given that Twitter is being used normally from specific fragments of the society such as the younger population of the European societies and/or politicians, experts and journalists, they allow us to explore a big number of observations which is usually impossible via more traditionally methodological approaches due to cost reasons and time limitations. In addition, Twitter data permits us to estimate usually neglected parts of the society that are not participated in traditional surveys or do not have the option to become part of a social science survey. In this sense, I argue that data stemming from Twitter are more representative samples of the EU population whereas, at the same time, picture a larger fragment of the EU member-state societies that participate actively in the construction of the European politics discourse.

My future research agenda focuses on the examination of the evolution of the EU as a global security actor through a large-scale multi-source study based on the use of advanced CSS approaches. Drawing on a vast amount of textual data from a rich variety of sources and exploiting a wealth of research instruments stemming from CSS methods (more specifically, automatic content analysis, topic modeling, named-entity recognition and sentiment techniques), I will try to identify the dominant discourse and the major events that have played a significant role in the formulation and determination of EU security and defence policy. In addition, I will attempt to recognize the dominant beliefs and the correlated discourse the citizens in foreign countries (like the USA, Russia and China) have constructed with regard to the EU as a global security actor. To do so, I will create a large event database capturing events that happened in the timespan of the last twenty years and which are related to the phenomenon under study. All entities (people, organizations and locations) involved in these events as also as the sentiments and emotions expressed will also be captured and coded in a knowledge network facilitating the exploration of the perceptions concerning the EU as a global security actor.

(*) The MWPBlog is a platform for MW Fellows to address scholarly topics and comment on current affairs. The thoughts expressed in the posts represent solely the views of the posting Fellows and not of the Max Weber Programme