View article

We study the effect of social media on xenophobic attitudes in Russia. We build a model where social media increases the likelihood of meeting like-minded people locally and in other cities, and where online interactions can be persuasive. We show that social media increases the share of individuals holding extreme opinions, but it may also increase the share of people hiding these opinions, which calls for proper measurement of attitudes. Empirically, we confirm these predictions by combining data from a survey experiment with data on hate crimes, and exploiting quasi-exogenous variation in city-level social media penetration in Russia. We find that higher city-level social media exposure: i) increased the share of individuals holding xenophobic attitudes (consistent with a persuasion mechanism); ii) reduced people’s willingness to openly express xenophobia (consistent with a social image mechanism); iii) led to more hate crimes in cities with higher pre-existing levels of nationalism (consistent with a mechanism of connecting like-minded people locally, thus facilitating the coordination of collective action).