According to a popular story in the New York Times, the general use of Facebook 鈥溾�� racist attacks in Germany. The story, based on what they describe as a 鈥溾�� paper by two University of Warwick PhD students, says that in towns and cities where Facebook is used more than average, there are more attacks on refugees and immigrants. Both the story and the paper say that this is more than just a correlation 鈥� the Facebook use itself causes an increase in the number of attacks.
But both the study and the New York Times鈥檚 reporting of it have . Rather than looking more generally at whether Facebook use drives anti-immigrant attacks, the study looked at something more specific 鈥� posts and likes on two Facebook pages, that of the chocolate spread Nutella, and that of the German hard-right anti-immigration party Alternative f眉r Deutschland (AfD).
The study found a perhaps unsurprising link that areas where more people like anti-immigrant Facebook pages are more likely to have attacks on immigrants. But the study also found that the pattern of likes and shares of the Nutella page also seem to be linked to immigrant attacks. The researchers say this shows that it is Facebook use in general 鈥� not specifically AfD page activity 鈥� that correlates with these attacks.
Advertisement
Sticky issue
However, the Nutella comparison has problems. The page had only 22,000 likes from people in Germany, and there are areas of the country where no one has liked the Nutella page. This makes using Nutella a less-than-ideal proxy for general Facebook use 鈥� the data is too sparse.
What鈥檚 more, at times when Facebook usage and immigrant attacks both rise, it is hard to detect causality. 鈥淛ust because the two things spike at the same time, doesn鈥檛 mean that Facebook is fuelling the incidents,鈥� says Jennifer Rogers, a statistician at the University of Oxford.
In one graph in the study, Facebook activity spikes after a rise in attacks 鈥� suggesting that attacks may make Facebook activity more likely, not the other way round, says Rogers.
Equally, some third factor could be causing them both. One large spike in Facebook activity and racist attacks, for instance, coincided with the 2015 New Year鈥檚 Eve attacks in Cologne, in which a group of immigrants attacked women. Those attacks could easily have sparked the Facebook activity as well as revenge attacks on immigrants.
Plausible effects
That鈥檚 not to say that Facebook doesn鈥檛 influence racist attacks or any other activity, says Charlie Beckett, who studies media and communications at the London School of Economics. 鈥淚 don鈥檛 think it鈥檚 implausible that it has an effect, that it stimulates people,鈥� he says. 鈥淭here is evidence that media has an amplifying effect. People in their emotionally heightened online world might translate that into action.鈥�
One way is through giving 鈥減ermission鈥� 鈥� in the Arab Spring, he says, people found out via social media that everyone else around them wanted to overthrow their government, and that gave them social permission to say it. It also can be used to organise: again in the Arab Spring, people organised the protest in Tahrir Square in Cairo, Egypt, via Facebook and Twitter. It is easy to imagine people organising attacks on immigrants using the AfD Facebook page.
But neither the study, which does not appear to be peer-reviewed, nor the New York Times story shows this. And the more eye-catching claims in the story 鈥� that Facebook use leads to up to a 50 per cent increase in attacks 鈥� are implausible. If there is a causal link, it is probably much weaker and more confused than that.
SSRN
听
