A study conducted by Karsten Müller and Carlo Schwarz - both of the University of Warwick - of the various attacks against refugees and immigrants in Germany is deeply linked with the parallel frequency of use of Facebook, with increase in anti-immigrant posts being predictive of violence against such groups.
Their primary theory here was that before manifesting in real-life as violent hate-crimes, waves of "right-wing anti-refugee sentiment" would ripple countrywide via word of mouth, mass media and social media. Their study centered in on the latter, correlating spikes in pertinent Facebook activity with tangible anti-refugee crimes.
In order to monitor right-wing activity specifically, Müller and Schwarz looked at the Facebook page of the Alternative for Germany political party, currently the most popular anti-immigration political group in the country. The page was not reputed for its wholesome comment threads. As a point of reference for general, overall use of the platform, Müller and Schwarz kept tabs on the public German page of Nutella.
Having classified the content they were observing based on time and location, the two were able to isolate anti-refugee sentiment in the overarching social media pattern of use, giving them rather unambiguous findings: there's a large, disproportionate spike - estimated to be roughly by 13% - in anti-refugee hate crimes such as assault and arson during periods of higher than usual anti-refugee sentiments online.
The researchers were careful to communicate that social media itself was not the root cause of these crimes, saying:
"... hate crimes are likely to have many fundamental drivers; local differences in xenophobic ideology or a higher salience of immigrants are only two obvious examples. Rather, our argument is that social media can act as a propagating mechanism for the flare-up of hateful sentiments. Taken together, the evidence we present suggests that quasi-random shifts in the local population’s exposure to such sentiments on social media can magnify their effect on refugee attacks."
Of course, directly correlating two proportional numbers isn't ever a reliable way to draw parallels for anything, leave alone an extremely complex issue such as immigration which has several equally complex factors influencing it. As a result, the researchers made sure to provide evidence of such connections through the process of elimination.
The first potential fallacy could be described as thus: what if hate crimes are simply more likely in areas with greater population density and therefore more social media use? There would, after all, be no point in drawing a connection between these two metrics.
This is where the Nutella control group came in: it gave the researchers a data point that is as politically agnostic as possible, so usage patterns can be more easily analyzed so as to give room for rhythms that accompany seasons and holidays. Consequently, any deviation from these norms among users would be noticeable.
Secondly, the researchers focused their attention not on all of Germany at the same time, but in smaller towns and areas with relatively lower levels of social media engagement. This would make looking for anti-refugee content a lot easier given the different demographic and simply by virtue of having far fewer numbers to deal with.
Another unprecedented factor was Facebook and general internet outages:
"The effect of refugee posts on hate crimes essentially vanishes in weeks of major Facebook outages.”
Moreover, these hate-crimes tend to focus on refugees alone and not other groups that are often targeted by fundamentalists; negative sentiment against the Jewish community, for instance, was not accompanied by hate-crimes against them.
Lastly, the researchers also discovered that while coverage of refugee-related issues on major news outlets did drive engagement among people in the form of protests, for instance, it did not act as a precursor to acts of violence in the same way Facebook did.
This study makes considerable headway into plotting the use of social media to further stoke anti-immigrant sentiments, and only serves to further emphasize the very real, very dangerous consequences social media has on geopolitics. But whether or not it actually winds up holding any party accountable for indirectly spreading violence is another matter altogether.