Anti-European protests in Ukraine

Studies of Anti-European protests (that often overlap with Pro-Russian protests) in Ukraine are scarce. There were at least three major reasons for this: too dangerous, too difficult, too stigmatised. The impact of each varied in the course of time. Furthermore, the subject remains quite sensitive making retrospective surveys impossible. It is hard to imagine someone walking around reclaimed territories asking questions like “so, did you participate in Anti-European movements last year?”. Not to mention that the research community in Ukraine is in hassle. Large-scale surveys in Crimea and Eastern regions are conventionally condemned after GFK surveyed Crimean population (the reaction of Sociological Association of Ukraine is here). 

Of course, there are some exceptions. For instance, a brilliant qualitative study during the emergence of the pro-European and anti-European movements in Ukraine was conducted by PS.Lab. But to my knowledge it is written only in Russian. And, as any other qualitative study, it spots important phenomena but says nothing about the magnitude. Another important research has been done by the CSLR (the report is in Ukrainian language). However, the crude number of protests says little about the interactions between the participants, their motivation, mobilisation, and values.

Having said this, I think there is a way to fill at least a small part of the gap in knowledge we have now. Looking at online conversations and interactions among anti-European/pro-Russian activists it’s possible to spot the dynamic of the political mobilisation, to see the real-life invention of the group identities, to study what kind of obstacles people faced in their attempts to build a dialog.

So far I have made only few steps to explore this. First of all, I have collected the data from the “Antimaidan”  Facebook page that was launched as a response to its pro-European vis-a-vis “EuroMaidan”. The data cover daily online interactions among activists for February – May 2014.

Some simple descriptives can be a starter. This figure shows only March-April 2014 (I do not show all the data here for technical reasons).

I like using this measure of Comments-to-Shares ratio as a metric of how often people were engaged into a dialog when compared to simple spreading. I did the same calculations for the EuroMaidan page, and it seems that Anti-European page witnessed quite more conversations. Why is that? I guess the most accurate explanation would be the content of posts. Pro-European page signalled a lot of urgent requests for actions. People had to come at some place, bring resources, help others – no time for talks. Whereas at Anti-Maidan page people were engaged in real mobilisation through sharing messages and debating them.

Here I try to look at this by measuring reciprocity and the content of ties.The following graph includes several posts published during at the very beginning of the existence of the Anti-Maidan Facebook page (late February 2014).

Large red nodes numbered 1 to 7 are posts. Blue nodes are people connected to them by reading. Edges between blue dots = conversations between people. The point of an edge shows the direction. It may be reciprocal, may be not.

Yellow edge – they are fighting. Pink edge – complementing each other

You may see a number of isolated nodes – people read the posts but never talked to each other.

As you see  there were a lot of fights between people. So this group – at least at the beginning – was really about people with different views arguing about their believes. And more detailed investigation of their debates may show when and how the threshold was passed and the Pro-Russian/Anti-European majority excluded their opponents.

The last technique that may be handy in studying the developing of this group is the dynamic visualisation of the network.

This video features people (blue nodes) reading posts (red nodes). Some new readers appear in time following new topics, some of them disappear. But a big proportion of readers are quite stable in the course of time. It would be very interesting to investigate whether the views of these stable readers changed over time due to the overall network developments.

Of course, this post hasn’t solved any puzzles, but I think it shows at least that the process of online mobilisation among Anti-Maidan activists was very much dynamic. And this line of research may be very fruitful.

Advertisements

Dynamic networks visualisations

Hail to the Gephi!

Dynamic networks visualisations in Gephi provide a nice leverage in analysing network data.

This is an example of tweet mentions of Poroshenko (Ukrainian poser-tycoon president) for April 10 – April 16. The data were collected with Socioviz.  TMs are not very useful per se. No friendship connection, unobserved variables, endogeneity and all other nightmares of a researcher.

However, dynamic analysis may do a trick. I invite all to read this brilliant paper by Moody and McFarland about it.

Here is a mere 20 second video of the Poroshenko’s mentions. There are almost 8,000 nodes, so the graph is a bit messy. But you still can see how cliques emerge over time.

 

 

 

Does Facebook generate obstacles for social capital? Some thoughts on the Ukrainian revolution case

Ukrainian revolution EuroMaidan has been considered by many as a booster for social capital. Online activism, allegedly, is a main contributor. As the story goes, people received and spread information engaging in dialogs with anonymous others, thus increasing reciprocal trust and coordination. However, as we know, research is quite sceptical about Facebook and activism. The question whether Ukrainians used Facebook to facilitate social capital online still begs an answer. In my opinion, the key here is social network properties. Why? Lets break an answer in a few simple steps: (1) Social trust and reciprocity are embedded in social networks. Thus, if we cannot observe social network than the overall question of social capital has no point. It is simply not possible to build social capital if there is no social fundament for it (networks).  (2) And how do we observe networks? By looking at connections of people and measuring their properties such as overall amount of connections vs all possible theoretical connections; shortest path between 2 people; or the likelihood to be on the path between other people…in other words –  network properties. So, what can be done to investigate social networks of Ukrainian activists? During the last year I had systematically collected the data of comments from the Facebook page “EuroMaidan” that was developed for the coordination and mobilisation of activists. Netviz was a very helpful tool for this task (many thanks to Bernhard Reider for his job). I collected 946,776 comments to 27,458 posts from November 2013 to May 2014. I collected them on the weekly basis making 25 networks. Then I divided them in four periods common for the Ukrainian researchers: a brief period of Pro-European protests; the so-called Revolution of Dignity; Violent clashes on Maidan square; and everything after the end of the revolution when the president fled the country. No real friendship can be observed among readers of “EuroMaidan” page. The only one thing we can see is the affiliation by posts. In other words, we can have 2-mode networks where readers are connected by reading the same post. These 2-mode networks are converted to 1-mode networks (I used Pajek for all data handling). What can we say about the density and clusterization of these networks? I also generated random graphs for the same periods to compare my empirical values with the simulated ones. Screen Shot 2015-04-12 at 1.49.09 PM Screen Shot 2015-04-12 at 1.48.58 PM What can be seen from these two figures is that at the very start of the revolution both density and clusterization were almost at the same rates as random. Meaning that social networks of online activism were not structured yet. This gives me a confidence to think that the initial recruitment to the Facebook page was not driven by the prior friendships. Otherwise we would have seen something opposite – friends would have been commenting the same posts increasing density. But there is also another important finding. Both density and clustering were quite stagnant during the revolution and increased only soon after. Why is that? Again, it is hard to imagine that the likelihood of a comment is driven by the mobilisation of friends. I would suggest that this have something to do with the content of posts. The posts during the revolution were about urgent requests for help and coordination (where to bring supplies, how to escape police, who needs help etc). The next figure illustrates it quite well. This is the ratio between an amount of comments and shares per day during the all observed period. Screen Shot 2015-04-12 at 2.01.37 PM At the very beginning this ratio was below 1. And then, after the end of the revolution, it grew to the stable levels of greater than 1. Meaning that during the revolution people just shared urgent important requests. It is only after the revolution the content of posts changed from urgent requests to more speculative news, sometimes about geographically remote places (such as which candidate is better for the upcoming elections, or what are the political consequences of pro-Russian activities in Crimea). Being relieved from hundreds of urgent requests during a day, activists received more time to comment posts and each other. So what is the take-away of this post? I hope to prove that online activists did not have an opportunity to generate social capital during the revolution. They had to react urgently and share news, and their comments were not dependent on the commenting of other people. The latter point is the key. When people comment on their own without being affected by other activists, this means that there is no real social embeddedness. Overall, Facebook page served a great purpose of being informative source. But it hardly was a space for social interactions. What was the main obstacle for social capital emergence? Ironically, it was the quantity of posts. The more posts were produced per day, the fewer comments and, consequently, dialogs were observed.

Facebook and revolutions. Video-Conference for the CIUS, 2014

Online social networks are known to shape revolutions or mass riots in the 21 century. Ukrainian mass protests in the end of 2013 (also known as EuroMaidan revolution) are not an exception.
Some of the first attempts to speculate about the role of Facebook in facilitating social mobilisation among Ukrainians were presented during the conference organised by the Canadian Institute of Ukrainian Studies.

 

Hi! My name is Tymofii Brik, I am a PhD in Social Science. This blog communicates my empirical findings and speculations on the subject of social activism online.
My email is tbrik@clio.uc3m.es