Social Media and Xenophobia: Evidence from Russia

How to cite

Bursztyn, Leonardo; Petrova, Maria; Enikolopov, Ruben; Egorov, Georgy. 2018. "Social Media and Xenophobia: Evidence from Russia." AEA RCT Registry. AEARCTR-0003066. https://doi.org/10.1257/rct.3066-1.0

Download: BibTeX · RIS — for reference managers without a browser translator. Zotero and most others pick up this page's citation metadata directly.

Leonardo Bursztyn* , Maria Petrova , Ruben Enikolopov , Georgy Egorov

* Primary investigator

Registry ID
AEARCTR-0003066
DOI
https://doi.org/10.1257/rct.3066-1.0
First registered
Last updated
Status
In development
Trial dates
2017-05-20 – 2019-05-01
Countries
Private
JEL codes
D7, D8
Secondary IDs
N/A
Registry record
http://www.socialscienceregistry.org/trials/3066

Abstract

In recent years, a large increase in expression of hate has been documented (for numbers on the US, see Center for the Study of Hate and Extremism, 2017). Moreover, in recent elections, nationalistic platforms have also received considerable support (e.g., Donald Trump, Marine Le Pen and Brexit). In this project, we plan to analyze the causal effect of access to social media on animosity toward specific groups of population (e.g., ethnic groups) and hate crime in Russia.

Social media reduces the cost of coordination, which might be particularly relevant for illegal/stigmatized activities, such as hate crime. Social media makes it easier to find like-minded people and might reduce the cost of asking or exposing one’s self. Social media might also influence or reinforce people’s opinions: tolerant individuals might be more exposed to intolerant views, while intolerant individuals might end up in an “echo chamber”.

Exploiting a natural experiment that generated plausibly exogenous variation in social media access at the city level in Russia, we analyze whether in cities with a high baseline level of private support for nationalistic ideas, access to social media increases hate crimes and the prevalence of hostile attitudes against ethnic minorities. We have already found a positive impact of social media access on the incidence of hate crimes against ethnic minorities.

The survey is implemented as a part of this project, with the aim of understanding better the channels through which social media affects hostile attitudes against ethnic minorities. In particular, we plan to implement a survey experiment in several cities in Russia to measure social stigma associated with xenophobic attitudes by asking survey participants about their attitudes toward people of other ethnicity and migrants both directly and indirectly through a list experiment (which allows eliciting the same attitudes on average without revealing preferences of individual respondents). The difference between the share of respondents that have negative attitudes toward people of other ethnicity and migrants as measured through the list experiment (i.e. in an indirect way) and those that admit such attitudes in response to direct questions will be used as a measure of social stigma, associated with such attitudes (under the assumption that social stigma makes people less likely to reveal such attitudes even in the context of anonymous survey. This measure will be used in a city-level analysis of the effect of social media on both this measure of social stigma and their “implicit” attitudes toward people of other ethnicity and migrants revealed indirectly through the list experiment.

The project will also consider the heterogeneity of the main effect by gender. An analysis of over 100 VK communities associated with hate attitudes indicates that about 3/4 of their members are male, so we expect to observe stronger effects for male respondents.

Intervention

List experiment to reveal social stigma associated with xenophobic attitudes, and also a measure of xenophobic attitudes with a higher degree of plausible deniability.

Experimental design

In the list experiment framework we will randomize whether a statement about xenophobic attitudes is included or not in the list of statements for which respondents asked to indicate the number of statements with which they agree. Respondents who end up in the treatment group in which this question is not included in the list of statements are subsequently asked this questions directly.

Design details

Randomization method
Randomization done by a computer of the survey firm.
Randomization unit
Individual.
Clusters
100 cities.
Observations
5,000. Importantly, we have run a pilot of 1,000 observations from 20 cities, in May 2018. The goal was to test the instruments and logistical implementation. The instrument used is nearly identical to the one to be used in the next data collection. To gain power, we will pool the observations from both samples. As a robustness check, we will also show the results dropping the observations from the pilot.
Arms
2500/2500
Primary outcomes
- City-level xenophobic attitudes; - City-level social stigma associated with xenophobic attitudes.
Secondary outcomes
- City-level VK penetration; - Pre-existing nationalism.

Cited by

No citation data available for this registration yet.

Keywords