Public debates on the language situation in Luxembourg and especially about the status of Luxembourgish have been quite vivid during the past months and they were mostly triggered by the registration of two petitions (Pétition 698 and Pétition 725; see our previous analyses of the geographical distribution and the family names of their endorsers). Both led also to controversial discussions in social media. One Facebook page became the main hub for these discussions: This page was created in the context of the referendum held in Luxembourg in 2015, which, among other questions, asked Luxembourgish citizens whether foreign residents should be granted voting rights. became a “content farm” disseminating arguments against such possibility – hence its naming (nee means “no” in Luxembourgish). Although the posts on the Facebook page focus heavily on aspects of language(s), there is room for related topics such as migration, economic growth, and fears associated with social change in the highly diverse context of Luxembourg. This Facebook page remained active after the referendum and gained new momentum with the appearance of Pétition 698.

This Pétition about the official status of the Luxembourgish language will be discussed in the country’s Chamber of Deputies on January 16 2017 (together with Pétition 725). Political parties will therefore have to take positions in respect to Luxembourg’s national language.

In this context, this blog entry collates the activity in with the one in the Facebook pages of the main Luxembourgish political parties. By doing so, we not only grasp the dimension of the explicit Facebook discussion related to Pétition 698, but also observe how users interested in these topics engage with the Luxembourgish political spectrum. For our analysis, we collected all the activity on the Facebook pages stretching from September 1 to November 30 2016 (Pétition 698 was registered in the Chamber on August 16 2016).

The following table shows the Facebook activity hosted in and in the pages of seven Luxembourgish political parties during the selected period (the Piratepartei is currently the only party without representation in the Chamber of Deputies).

Facebook pageLike count*Number of PostsUnique users***Percentage of users also active on Nee2015.luOverall sharesOverall reactionsOverall commentsEngagement
Déi Lénk8973159118%4869721201578
Piratepartei Lëtzebuerg151**206611%6516110236
Déi Gréng4152823717%9542438557

* Official Facebook like counter as retrieved on December 7 2016. Facebook pages may show a higher counter due to privacy settings of users liking the page.
** Official Facebook like counter as retrieved on December 13 2016.
*** Number of users liking or commenting one or more posts throughout the period (it does not consider sharing).

At first sight, we can already notice the dimension of the activity that took place in This page hosted more user engagement than the sum of all the political parties. This is especially significant since only published 30 posts, in contrast, for example, to the 172 posts by the CSV. Thus, Pétition 698, which was the underlying main topic of discussion in during the selected period, attracted a huge amount of attention to this page. As for the political parties, if we compare their current representation in the Chamber with the engagement their Facebook pages receive, we can observe rather low engagement for the LSAP and “déi gréng” and high involvement with “Déi Lénk” and, especially, the ADR.

The number of unique users is an important aspect because it tells us how many different people engaged with a page. In our case, this variable further confirms the enormous activity that took place in Luxembourgish political parties, in turn, attracted many fewer users. All these people, though, might had participated in other pages as well (Facebook users can participate on as many pages they have “liked”). The table above shows the percentage of users of each page that were also active on during the selected period. As we can see, a majority of the users of the ADR page were also active in, a much higher proportion than users of any other political party.

The following network further illustrates the participation overlaps between these Facebook pages. It is a visualization of the data in the table above, in which white nodes represent Facebook users and coloured nodes Facebook pages. The size of a node correlates with its engagement (reactions and comments, but not shares). In this way, the 30 posts in appear merged into one single node whose size relates to their total engagement. Besides nodes, the graph is full of links between users and pages. What do these indicate? Each link means that a user commented or reacted on a post from that page. When a user commented or reacted more than once with one or more posts of that page, the resulting link becomes thicker (although this is difficult to appreciate in this graph). Moreover, links affect the location of each node. Simply speaking, two connected nodes will attract each other and will appear closer than unconnected nodes.

Same graph as interactive network (though with no correlation between page node size and total engagement)

In short, such visualization allows to observe whether during the selected period a user engaged with more than one page. If the users of one page only participated in that page, an isolated cluster would appear. This doesn’t happen in our graph, though. Instead, we see users connecting with several pages. For example, let’s look for a moment to the LSAP and the DP. Above both nodes, we can see what we can call an “exclusive audience”, that is, users that during the selected period only participated in that page. Between the two, though, we can also distinguish a small cluster of users which, actually, participated in both pages.

Now, let’s look at what happened with As discussed above, this page had lots of activity by lots of different people, generating an enormous “bubble” around it. However, besides its exclusive audience, we can observe that a considerable part of its audience also reaches to other pages. This means that, during the selected period, there were Facebook users that engaged both with and with the political parties. Above all, and this is important, they engaged with one political party, the ADR. This explains why the ADR is the closest page to

Interpreting this graph requires some precautions. We must keep in mind that a user disliking a post is represented in the same way as one who liked the same post. This should prevent us from inferring that more engagement equals to more popularity or that people engaging with any page are hence their supporters. To confirm this, an additional qualitative layer would be needed. However, we know that, as a result of the personalization enabled by social media, it is more likely that users become exposed to content that supports their views rather than one that may challenge them. In this way, it is more plausible that engagement relates to users’ alignment – though we can’t be certain about it.

Conclusions concerning Facebook activity

  • All political parties seem to have their own exclusive audience.
  • Many users, however, engage with more than one party.
  • People seem to be more interested in than in the activity of political parties.
  • There is a considerable overlap between users active in and those active in the political parties.
  • The political party with most overlap with is the ADR.

Är Commentairen