Protests against austerity continue to roil parts of Europe, most recently in Brussels earlier this month when 100,000 people took to the streets and police deployed water cannons.
This type of public and occasionally violent demonstration has been taking place across Europe since the beginning of the sovereign debt crisis as leaders decided cutting spending was the best way to deal with the region’s mountain of debt. The resulting cuts in benefits and education spending alongside higher taxes and freezes in public sector salaries helped galvanize many Europeans onto the streets in protest.
But beyond the immediate impacts of austerity and spiking unemployment rates, what other factors shaped the public’s reaction? And what mobilized people to demonstrate in ever-rising numbers across Europe?
A new study that I published in the International Journal of Communication helps identify those factors and signals that they go well beyond the usual suspects of unemployment, earnings and social demographics such as population density.
In fact, my research shows that one of the most basic forms of modern communication and simplest uses of a smartphone, the text message, was crucial to mobilizing anti-austerity efforts across the European Union. Through a combination of statistics and network analysis, my study paints a picture of what drove the protests and who was behind them.
The study began with national-level data on 27 EU member countries (all but Malta). These data were used to look at the relationships between the prevailing economic conditions over the course of the economic crisis, from 2007 through 2012.
The analyses started with simple correlations, which found some relationships between youth unemployment and protest activity that got stronger over time, as well as very weak but positive correlations between protests and the number of texts sent.
Where it started to get much more interesting, though, was when I controlled for the influence other variables were having on protests, such as democracy levels and population density. This was done with what is known as a regression modeling, which is similar to correlations but allows researchers to account for the influence of other variables, including time, on an outcome. Unlike correlations, regressions can get closer to (but still not perfectly identify) causal-type relationships.
What I found in the first regression model was that the volume of text messages sent in all countries was a positive predictor of protest activity, as were youth unemployment rates. What this means is that as the total number of text messages sent went up in one year, protest activity increased in the next year. The same was true for youth unemployment rates. More protests followed rising levels of joblessness among young people.
Combining youth unemployment and texting
My statistical analysis showed these relationships were not a result of mere chance. The regression models took into account a range of other related factors that perhaps could have also explained the increases in protest, such as internet diffusion and income levels. But even when taking into account those factors, youth unemployment and texting levels remained significant.
Even more interesting, though, was that youth joblessness showed this same relationship only in an overall model, and not in year-by-year regressions. That is, it showed up only in figures for the entire series of countries and years. Of the seven factors that were modeled, only texting activity remained positive and statistically significant from 2009 onward in explaining protest movement.
So to try to make the most of all these findings, I then combined the volume of texts with the youth unemployment rate to determine the collective impact of these two factors at the same time. Looking at what is called an “interaction term” showed that the combination of these two factors had an even stronger relationship to higher levels of protest.
In other words, while rising youth unemployment appeared to be a primary catalyst prompting people to protest, it was the combination of youth unemployment and the increased frequency of texting on mobile phones that was related to demonstrations swelling in both frequency and intensity over this six-year time period. The graph at right shows the convergence of these factors.
Modeling the communication
So what were people communicating through their mobile devices? And which mobile phone users were most pivotal to coordinating these anti-austerity protests?
While it wasn’t possible to access the content of all the texts zooming back and forth over this period, for obvious reasons, I was able to examine more than three million public tweets that included relevant keywords such as “austerity,” “euro” and “crisis” – most of which were sent from mobile devices and all during the last two months of 2012. The tweets serve as a useful stand-in for the content and users of texts because tweets are more visible and user names (unlike phone numbers) are publicly accessible.
Using BU’s Twitter Collection and Analysis Toolkit (BU-TCAT), I applied several algorithms to sort the Twitter data to reveal underlying network structures that shaped the flow of information in the online community. One of the resulting graphs visualized the most prominent hashtags used, while another identified the most influential users discussing the euro crisis on Twitter at the time.
The first of these graphs, the figure below, illustrates the mix and interconnection of topics related to the euro crisis. In general, this concept map provides a level of understanding about what was being communicated in this public space and how those conceptual interactions occurred, without specific coordination or hierarchy, by Twitter users.
When looking at this graph of co-occuring hashtags, to understand its meaning, imagine if you gave 1.5 million people an orange and asked them to describe its features. Eventually certain dominant descriptive keywords and linkages between words would emerge, such as “juicy,” “round,” “sweet,” “orange” and so on. Now, when looking at this co-hashtag graph, imagine that the 1.5 million users that tweeted weren’t given an orange but instead a chance to publicly express their opinions on austerity policies. The most prominent nodes in the co-hashtag graph identify which concepts users applied most to describe their opinions on austerity policies.
You can click here to see a dynamic version of the graph with more detail.
A second graph illustrates how there were relatively few connections among the diverse stakeholders operating in a network where users were surprisingly sparsely interconnected with one another. More specifically, what this means is that there were more than three million tweets in this data set from just over 1.5 million users, but that in the period of these two months, there were just 227 links between the 409 most active users.
To clarify further, this data set was summarized into the leading 409 figures that were tweeting about the euro crisis and austerity on Twitter, but there were just 227 instances in which they mentioned one another using the @username reference. You can see a dynamic version of this here.
Clearly, there was a large number of tweets and a broad base of users. Still, this analysis showed that user interaction was irregular, generally uncoordinated and centered around a relatively small number of users. Additional research identified these users as mostly belonging to a community of Spanish journalists, activists and media organizations, along with satirical or pseudo accounts (such as a deceased politician).
This study signals that these ostensibly open spaces for communication often morph into something else. Rather than being a connected group of equal participants, the Twitterverse was full of disconnected posts and users where only a few acted as influential gatekeepers to share content.
Smartphones and simple tech rules
Overall, this study found that it was how people used mobile media and not just the pervasiveness of smartphones (which in itself was not statistically significant) or online access more broadly that were pivotal factors in how the public often erupted in protest to austerity measures.
In short, the simplest usage of the smartphone proved the most crucial factor – when combined with youth unemployment – in explaining where and when people went to the streets to oppose austerity measures.
This is unique and different from how protesters in the Arab Spring or other (reportedly) technologically inspired revolutions have been considered. Whereas those protests also engaged mobile phones, texting and social media, they were operating in a different cultural and political climate seeking regime – not policy – change. The results, of course, were also vastly different in that governmental transitions in Europe have been more orderly and programmatic, at a minimum, and that austerity politics have continued.
Sometimes, the simplest answers can be the most profound. In this case, the evidence suggests that the public erupting in protest against austerity measures was being driven, above all, by youth unemployment and texting.
In the study of emerging media and politics, newness is not often nearly as important as “embeddedness” – in this case, the device we all carry in our pockets, smartphone or not. Simple matters – and, as suggested here, sometimes short (message services) matters too.